element is appropriate >here, but the periods supposedly added for the sake of symmetry (?) >completely ruins any symmetry instead. ... .............. >Maybe we need an H.T. Emily Webster set of guidelines regarding how >(not) to write good (or bad) HTML code. ;-) BILL MINNICK P.S: Good Idea; please produce a strawman draft for review and criticism.

<pre>The genweb mailing list began in September 1994 with these messages.


From bgedsu02.nho.hydro.com!birger@netcomsv.netcom.com Thu Sep  1 08:34:30 1994
Date: Thu, 1 Sep 1994 17:09:50 --100
To: ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        pakers@netcom.uucp, hstoyan@faui80.informatik.uni-erlangen.de,
        shadow@mindspring.com
Subject: GenWeb
Cc: birger@sdata.no


First of all: Reply to me as birger@sdata.no, not at this address....
I just wanted to get this mail out. If I waited until I got back to
the office, I would have forgotten the whole thing.....


I have had a quick tour around the various genealogy-on-WWW demos.
Before I found these, I started experimenting on my own.

My approach was to access the data directly from my LifeLines database.
That is, after all, where I edit the data, and having up-to-date
data available directly avoids a lot of hassle.

LifeLines is a PD UNIX-based genealogy database. It currently operates
directly on GedCom-like files. LifeLines' main strengths are:
- You get to write your data directly in gedcom format
- LL (LifeLines) only handles links between persons and relations.
- A report-generator programming language.

So I wrote a CGI (WWW gateway program) that runs a LL report,
returning a HTML page on the specified person.

The report currently only returns data on parents of a person, so you can only
follow links upwards to a persons ancestors. I don't want to do all the work
on completing the report program until I know what I want.

I currently see 2 ways to go:
- Either go all the way with my own report generator, having it generate
  HTML by itself
- or just write a report that outputs a gedcom entry for the person, and
  filter that output through Frode's ged2html.

I'll have to get ged2html, and then I'll decide.....

I have also asked LL's author to make some changes to allow easier running
of a report directly from supplied command-line arguments.

I only have 4 persons in this experimental base. To get a look at it,
look up my oldest daughter with the URL
http://www.vest.sdata.no/cgi-bin/ll-report/I3?Index

With my approach, I use the URL as follows:
Anything following the '?' will be supplied as command-line argument for
the script. This part will be subject to rewriting rules to conform to URL
standards. I use this for simple keywords. 'Index' is the only sensible
one at the moment.

Anything between /ll-search and the '?' will be put into an environment
variable, without rewriting, so I want to use this field for search
strings (for the planned 'Search' keyword), and for the index for the
'index' keyword.

Of course, the time I can spend on this is *very* limited.....


Birger Wathne
birger@sdata.no

From mindspring.com!shadow@netcomsv.netcom.com Thu Sep  1 09:55:40 1994
Date: Thu, 01 Sep 1994 12:46:22 -0500
To: abe@cc.purdue.edu
Subject: Re: GenWeb 
Cc: ghoffman@ucsd.edu, frode@ifi.unit.no, pakers@netcom.uucp,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        birger@sdata.no


>>http://www.vest.sdata.no/cgi-bin/ll-report/I3?Index
>
>This didn't work for me.  I got the message:
>
>   The requested URL /cgi-bin/ll-report/I3 was not found on this server.
>
>(No, I didn't mis-type the URL; I cut it from one window and pasted
>it into a lynx command.  :-)

I ran into the same problem here. I made three attempts to connect with no luck.
**************************************************************************
 Bill Spurlock                               5125-C BeverlyGlen Lane
 shadow@mindspring.com      Norcross, GA  
 (404) 368-8884                          30092
**************************************************************************
SPURLOCK FAMILY GENEALOGY HOMEPAGE :
 http://www.mindspring.com/users/shadow/shadow.html
**************************************************************************


From vic.cc.purdue.edu!abe@netcomsv.netcom.com Thu Sep  1 09:55:43 1994
To: birger@sdata.no
Cc: ghoffman@ucsd.edu, frode@ifi.unit.no, pakers@netcom.uucp,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        birger@sdata.no
Subject: Re: GenWeb 
Date: Thu, 01 Sep 94 10:51:30 -0500

Birger,

In message <9409011609.AA24773@bgedsu02.nho.hydro.com> you write:
>
>First of all: Reply to me as birger@sdata.no, not at this address....
>I just wanted to get this mail out. If I waited until I got back to
>the office, I would have forgotten the whole thing.....

Please put a Reply-to: line in your header so our mailers can
pick the right reply address automatically.

>I have had a quick tour around the various genealogy-on-WWW demos.
>Before I found these, I started experimenting on my own.
>
>My approach was to access the data directly from my LifeLines database.
>That is, after all, where I edit the data, and having up-to-date
>data available directly avoids a lot of hassle.

But that's not where others do their work.  I'm pretty well convinced
that using GEDCOM source is the correct method.  I've been able to
organize five different GEDCOM sources from different genealogical
DB programs via HTML/CGI without being limited to a proprietary DB
format.  If one were to choose a proprietary DB format, PAF's would
have to be first.

>So I wrote a CGI (WWW gateway program) that runs a LL report,
>returning a HTML page on the specified person.

It should follow all lineage links.  It should  allow regular
expression searches of the DB text.

>I currently see 2 ways to go:
>- Either go all the way with my own report generator, having it generate
>  HTML by itself
>- or just write a report that outputs a gedcom entry for the person, and
>  filter that output through Frode's ged2html.

I believe I have advanced Frode's ged2htm progarm many steps.  it
is more robust and uses a much faster method.

>I'll have to get ged2html, and then I'll decide.....

Get mine via anonymous frp from vic.cc.purdue.edu in pub/GED2WWW.tar.Z
or pub/GED2WWW.tar.gz.  (It's called ged2cgi, and the tar supplies
the CGI perl scripts as well.)

>I only have 4 persons in this experimental base. To get a look at it,
>look up my oldest daughter with the URL
>http://www.vest.sdata.no/cgi-bin/ll-report/I3?Index

I have one DB that has 9,000+ people in it.  That leads me to
believe my method will scale well.  Several earlier approaches,
based on Frode's initial work, didn't scale well at all.  You must
be very careful to avoid placing too much reliance on HTML links,
or having too many of them.  I think CGI is a better way to follow
lineage links.

Vic Abell

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Thu Sep  1 09:55:45 1994
To: birger@bgedsu02.nho.hydro.com (Birger Wathne)
Cc: ghoffman@ucsd.edu, frode@ifi.unit.no, pakers@netcom.uucp,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        birger@sdata.no
Subject: Re: GenWeb 
Date: Thu, 01 Sep 94 11:14:30 -0500


In message <9409011609.AA24773@bgedsu02.nho.hydro.com> you write:
>
>...
>
>I only have 4 persons in this experimental base. To get a look at it,
>look up my oldest daughter with the URL
>http://www.vest.sdata.no/cgi-bin/ll-report/I3?Index

This didn't work for me.  I got the message:

   The requested URL /cgi-bin/ll-report/I3 was not found on this server.

(No, I didn't mis-type the URL; I cut it from one window and pasted
it into a lynx command.  :-)

Vic

From freyr!pakers@netcomsv.netcom.com Thu Sep  1 22:53:34 1994
Subject: Re: GenWeb
To: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        pakers@netcom.uucp
Date: Thu, 1 Sep 1994 22:25:56 -0700 (PDT)

Hello all...
I may be missing something but I think that by going with any type
of conventional database you are throwing away some of the big ad-
vantages of HTML... the things that attract me to geneology with
HTML are the following...

1. It makes a perfect free form database on its own.
2. It is simple to add gifs (or other graphics formats)
   to the database.
3. It is simple to add audio files or even jpeg movies.
4. It is simple to add other documents to the database.
5. It is simple to add or change information.

After seeing all the information that my parents have collected
over 20 years research I just think that any type of conventional
database is too limiting... I want to be able to add the family
pictures to the database... maybe even audio tape or videotape
interviews (lots of diskspace of course!!)... scanned-in documents
like birth certificates, marriage licenses etc... maybe this is all
actually beyond what is normally considered to be a geneolgy data-
base but I think it is the best way to store the wide range of
information formats that you get in this type of research.

What do all of you think?

Phil
-- 

 Phillip L. Akers
 pakers@netcom.com
 freyr!pakers@netcomsv.netcom.com
 wa4dde@wa4dde.ampr.org (44.4.30.74)
 

From faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com Fri Sep  2 00:53:26 1994
Date: Fri, 2 Sep 1994 09:48:35 --100
From: Herbert Stoyan <faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com>
To: birger@sdata.no
Subject: GenWeb
Cc: ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        shadow@mindspring.com, pakers@netcom.uucp

Phil makes good points, I think.
But it does not follow that we could not use gedcom as basis.
We should define first what we really need.
We need:
1.access to a (bunch of) family tree which is represented in several gedcom 
  files.
  a) access via a person-directed url
  b) access via indices: alphabetical person index
                         time sorted person index
                         other dynamically generated indices
2.enlarging this locally with gifs,audio files,movies,other documents
3.linking to other family trees on different computers
4.the thing should be easy to work with (people which don't work with gedcom
  should be integrable, etc.)
What else?

Did you see my http://faui80.informatik.uni-erlangen.de/html/ww-person.html
? Herbert

From faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com Fri Sep  2 02:15:23 1994
Date: Fri, 2 Sep 1994 11:02:04 --100
From: Herbert Stoyan <faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com>
To: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: GenWeb

I missed:
Its not quite convenient to have 10000 of html files. Therefore, a system
of gedcom files would be a good solution for the management of the pure 
genealogical information.
I have some 30000 files at present...

From vest.sdata.no!Birger.Wathne@netcomsv.netcom.com Fri Sep  2 03:19:10 1994
Date: Fri, 2 Sep 94 11:49:50 +0200
To: abe@cc.purdue.edu
Subject: Re: GenWeb
Cc: ghoffman@ucsd.edu, frode@ifi.unit.no, pakers@netcom.uucp,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com


First of all an apology.

I was at a different location while writing the mail, and I did a mistake.
The URL is supposed to be http://www.vest.sdata.no/cgi-bin/ll-search/I3?Index

Now to comment on some of the responses I got....


Herbert Stoyan says:
>Interesting, Birger. I compile my gedcom into html via a self written 
>Lisp-program. How much is LifeLines? Where can I get it?
>Did you see my http://faui80.informatik.uni-erlangen.de/html/ww-person.html
>? There I have the problem of generating indices of people. 
>In which language is your cgi-program?

Lifelines is free. I currently run a beta version of 3.00. This version
may not be freely available. I know I saw an FTP site in one of the
genealogy WWW pages. You could also mail the author: thomas.wetmore@att.com

My CGI program is a shell script that starts a short lifelines session
to extract the wanted data.


Vic Abell says:

>Please put a Reply-to: line in your header so our mailers can
>pick the right reply address automatically.

I was at a customer site at the moment, and their mail systems does a
few levels of header rewriting of it's own, so I was unsure how the
header would end up anyway. I'm back at my office now, and my reply
address should be correct from now on.

>>My approach was to access the data directly from my LifeLines database.
>>That is, after all, where I edit the data, and having up-to-date
>>data available directly avoids a lot of hassle.
>
>But that's not where others do their work.  I'm pretty well convinced
>that using GEDCOM source is the correct method.  I've been able to
>organize five different GEDCOM sources from different genealogical
>DB programs via HTML/CGI without being limited to a proprietary DB
>format.  If one were to choose a proprietary DB format, PAF's would
>have to be first.

That's why I'm currently thinking along the lines of letting my report
output the info that it looks up in GEDCOM format rather than HTML.
Then I would only be able to use the GEDCOM to HTML filters already
developed.

I agree that we should standardize on GEDCOM to HTML conversions, but
there is no reason why the GEDCOM data has to be spread out in unmanageable
files, and not exported on the fly from the genealogy program I use
(since I happen to use one that runs on UNIX). This avoids data duplication
and data consitency problems. It removes any problems with invalid links,
as the links will be generated on the fly. But it becomes more difficult
to guarantee that an old index (e.g. in someone's hotlist) will refer
to the same person after reloading the database, merging databases, etc.
So I'm a little bit unsure about using lifelines' own index numbers
as the publicly available URL's.


>I have one DB that has 9,000+ people in it.  That leads me to
>believe my method will scale well.  Several earlier approaches,
>based on Frode's initial work, didn't scale well at all.  You must
>be very careful to avoid placing too much reliance on HTML links,
>or having too many of them.  I think CGI is a better way to follow
>lineage links.
Lucky you... I have a few thousand people on *paper*. And many
hours of typing ahead. I have a base with ~100 persons as well, and
I may migrate these into the test base one of these days.
I already use a CGI to a database report. This CGI returns a html page
that access other people through the same CGI. So I have nothing
floating around in HTML files. HTML output is generated directly from
the database on the fly. But with my current scheme there is a long-time
index consistency problem...


Phillip Akers says:
>
>Hello all...
>I may be missing something but I think that by going with any type
>of conventional database you are throwing away some of the big ad-
>vantages of HTML... the things that attract me to geneology with
>HTML are the following...
>
>1. It makes a perfect free form database on its own.
>2. It is simple to add gifs (or other graphics formats)
>   to the database.
>3. It is simple to add audio files or even jpeg movies.
>4. It is simple to add other documents to the database.
>5. It is simple to add or change information.
>
>After seeing all the information that my parents have collected
>over 20 years research I just think that any type of conventional
>database is too limiting... I want to be able to add the family
>pictures to the database... maybe even audio tape or videotape
>interviews (lots of diskspace of course!!)... scanned-in documents
>like birth certificates, marriage licenses etc... maybe this is all
>actually beyond what is normally considered to be a geneolgy data-
>base but I think it is the best way to store the wide range of
>information formats that you get in this type of research.
>
>What do all of you think?

That's why I'm converting from 'Brother's keeper' to lifelines.
Lifelines will only manage links between person and relationship entities
within GEDCOM. Apart from that, it's up to you to follow the GEDCOM
spec (and you can do so without any limitations from the program), or
invent your own tags. I browsed *very* quickly through the GEDCOM
standard yesterday, and I noticed there were tags for references to
external multimedia-oriented files. I want to have proper tags on all
information to make it easily to use it for purposes I don't yet
know I'll ned those data for :). But I also want a lot of external
info, like scanned images.


And then Herbert Stoyan is back with:

>Phil makes good points, I think.
>But it does not follow that we could not use gedcom as basis.
>We should define first what we really need.
>We need:
>1.access to a (bunch of) family tree which is represented in several gedcom 
>  files.
>  a) access via a person-directed url
This person-directed URL can be
	a) Defined by the person who owns the data
	b) On some standardized format
My system currently uses a), as there is no b).
My person-directed URL's currently rely on lifelines' internal indexing.
This should be avoided. I'll have to come up with some other kind of
index, but this is the easiest one for now (as this could be
seen as an early beta stage). I don't think we should enforce any
URL format. Some people may use flat files, some may use FTP access,
and others may use CGI gateways to various search engines.


>  b) access via indices: alphabetical person index
>                         time sorted person index
>                         other dynamically generated indices

>2.enlarging this locally with gifs,audio files,movies,other documents
GEDCOM has a tag for this, I think?

>3.linking to other family trees on different computers
How should this be expressed in GEDCOM? I must confess I don't know enough
about GEDCOM yet. I'll read that document one of these days...

>4.the thing should be easy to work with (people which don't work with gedcom
>  should be integrable, etc.)
I think we should standardize on the notion that whatever kind of data
people have, it should be possible to export it as GEDCOM. Most genealogy
programs are able to export at least the rudimentary data as GEDCOM.

>Its not quite convenient to have 10000 of html files. Therefore, a system
>of gedcom files would be a good solution for the management of the pure 
>genealogical information.
>I have some 30000 files at present...
>
I have 0 html files. Only my lifelines database :)


Birger

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Fri Sep  2 06:35:13 1994
To: freyr!pakers@netcomsv.uucp (Phillip Akers)
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        pakers@netcom.uucp
Subject: Re: GenWeb 
Date: Fri, 02 Sep 94 07:55:42 -0500


Phillip,

In message <9409020525.AA06840@freyr.uucp> you write:
>
>Hello all...
>I may be missing something but I think that by going with any type
>of conventional database you are throwing away some of the big ad-
>vantages of HTML... the things that attract me to geneology with
>HTML are the following...
>
>1. It makes a perfect free form database on its own.
>2. It is simple to add gifs (or other graphics formats)
>   to the database.
>3. It is simple to add audio files or even jpeg movies.
>4. It is simple to add other documents to the database.
>5. It is simple to add or change information.

HTML is not a data base manager and was not intended to be.  It
cannot cope with the large number of linkages that a genealogical
data base contains.

When I first started experimenting with Frode's ged2html converter
I used my 4,000 person data base as test data.  First I found that
the creation process generated 4,000 files, something few systems
could tolerate on any expanded scale.

Second I found that HTML linkage processing is inefficient and
cannot cope with the number of lineage links in a normal size
genealogical data base.  After modifying Fride's program to reduce
the number of files to a manageable size, but still using URLs to
effect the lineage links, I found that lynx couldn't load the main
index to my data base on a 25MHz 68040 NeXT cube in less than 30
minutes.

So HTML is not the complete answer to making genealogical data
available *efficiently* via WWW.  More work needs to be done on
the mundane chores of data organization and retrieval, and that's
what I have been doing.

>After seeing all the information that my parents have collected
>over 20 years research I just think that any type of conventional
>database is too limiting... I want to be able to add the family
>pictures to the database... maybe even audio tape or videotape
>interviews (lots of diskspace of course!!)... scanned-in documents
>like birth certificates, marriage licenses etc... maybe this is all
>actually beyond what is normally considered to be a geneolgy data-
>base but I think it is the best way to store the wide range of
>information formats that you get in this type of research.

Nothing we have done so far excludes audio, pictures or other
documentation forms.  The latest GEDCOM standard has room for all
of those.  I haven't had such data in the GEDCOM files I have been
working with, so I haven't had a chance to provide for them.  (I
did try to describe my work as experimental.)

Further development by someone who has the resources to deal with
the larger data requirements of sight and sound will clearly have
to deal with the data you mention.

Don't overlook the fact that the GEDCOM to HTML approach has the
virtue of accepting data in a common language, even though the data
has been recorded in a variety of tongues.

Vic

From faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com Fri Sep  2 06:35:15 1994
Date: Fri, 2 Sep 1994 15:32:42 --100
From: Herbert Stoyan <faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com>
Message-Id: <9409021332.AA24250@faui80.informatik.uni-erlangen.de>
To: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: GenWeb urls in gedcom


The question is then, 
a) how to encode links to people in gedcom files which differ from the current
   but are on the current computer,
b) how to encode links to images or else (at the current computer),
c) how to encode links to people external to the current computer.

For me, www for family trees makes only sense if the tree spreads over various
computers.


From ucsd.edu!ghoffman@netcomsv.netcom.com Fri Sep  2 09:36:36 1994
To: freyr!pakers@netcomsv.uucp, birger@sdata.no, frode@ifi.unit.no,
        shadow@mindspring.com, pakers@netcom.uucp, abe@vic.cc.purdue.edu,
        hstoyan@faui80.informatik.uni-erlangen.de
From: ucsd.edu!ghoffman@netcomsv.netcom.com (Gary Hoffman)
Date: Fri, 02 Sep 1994 09:29:42 PDT
Subject: Do we need a GenWeb list?

Hi, GenWebbers,

I am delighted with the recent discussions about HTML forms of genealogy
databases. Before we get too far here, we need to consider whether the
eight of us should form a new mailing list or whether we should continue
these discussions in GEDCOM-L. 

(I was recently bumped off of GEDCOM-L when my host was off line for a
couple of hours. I just got back on. I therefore thank those who included
me in their private traffic so that I could keep up during the GEDCOM-L
blackout.) 

After all, we are talking about genealogical data communication and GEDCOM
itself. My vote is to keep these WWW Genealogy discussions in GEDCOM-L  and
the rest of the world will follow. But if the majority of you believe
otherwise, then let's start a new list and keep this GenWeb discussion
moving.

FYI: I have proposed a presentation on GenWeb at the WWW Conference in
Chicago on October 17-19 and have invited Vic Abell to join me in the
presentation. The notification of acceptance was delayed and I still don't
know if they want this topic or not.  I am also considering a paper at
GenTech in January, but I will be taking the California Bar Exam in
February and I may need to study a bit for that ;-) Does anyone plan to
attend GenTech and discuss GenWeb there?

Cheers,
Gary


***************************************************************************
*Gary B. Hoffman, Computer/Language Lab Director e-mail: ghoffman@ucsd.edu*
*Graduate School of International Relations and Pacific Studies    (IR/PS)*
*University of California, San Diego (UCSD)          voice: (619) 534-7733*
*9500 Gilman Dr., La Jolla, CA 92093-0519 USA          fax: (619) 534-3939*
***************************************************************************


From faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com Fri Sep  2 10:35:32 1994
Date: Fri, 2 Sep 1994 19:21:24 --100
From: Herbert Stoyan <faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com>
To: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: GenWeb

We should compile a list of genealogical urls.

From faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com Fri Sep  2 10:35:35 1994
Date: Fri, 2 Sep 1994 19:07:55 --100
From: Herbert Stoyan <faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com>
To: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: gedcom-l and www

I have nothing against gedcom-l. I did not look in it, yet. I used soc.roots.

It would be a nice idea to present some of our systems in Chicago. I think,
I cannot be there. My system ww-person contains now ca 16000 people of german
nobility. I have ca 200 color images. The system descendants of charlemagne
is likewise big. Both systems are connected now.

I'm interested how many www-people do you believe are working (like Abe)
with Lynx? I'm removing all technical images just now out of the system.
in one hour the system will be up again.

For a technical paper we should have a solution proposal for the questions we
are discussing...

Separate issues are criteria of thrustworthness for data if we are willing to
add a new tree in the net of trees. 1 year ago there was a guy who had a big
mormon file of ascendants and descendants of Charlemagne. The descendants are
fine, but the other part was nonsense. One should have something like a checked
kernel and an attachment procedure.

From mindspring.com!shadow@netcomsv.netcom.com Fri Sep  2 10:35:37 1994
Date: Fri, 02 Sep 1994 12:58:25 -0500
To: freyr!pakers@netcomsv.uucp, birger@sdata.no, frode@ifi.unit.no,
        pakers@netcom.uucp, abe@vic.cc.purdue.edu,
        hstoyan@faui80.informatik.uni-erlangen.de, ghoffman@ucsd.edu
From: mindspring.com!shadow@netcomsv.netcom.com (Bill Spurlock)
Subject: Re: Do we need a GenWeb list?


>Hi, GenWebbers,
>
>I am delighted with the recent discussions about HTML forms of genealogy
>databases. Before we get too far here, we need to consider whether the
>eight of us should form a new mailing list or whether we should continue
>these discussions in GEDCOM-L. 

I, for one feel that the logical way to go here is with a seperate mailing
list. If my indications are correct, I can see the group growing rather
quickly and get the feeling that it might be better to to leave the GEDCOM-L
group to it's own things. I will go ahead and subscribe to the list just to
be on the safe side. I'm assuming that the method of getting on the list is
much the same as ROOTS-L?

I have also been considering doing a paper at Gen Tech this comming January
and have contacted them with a proposal, but yet to recieve a reply.

If you do present at the Chicago WWW Conference, let me know and I'll see if
I can't clear some time to come up for it.

I do agree that we've got a good thing started here, and I would like to
keep it going.

Bill

**************************************************************************
 Bill Spurlock                               5125-C BeverlyGlen Lane
 shadow@mindspring.com      Norcross, GA  
 (404) 368-8884                          30092
**************************************************************************
SPURLOCK FAMILY GENEALOGY HOMEPAGE :
 http://www.mindspring.com/users/shadow/shadow.html
**************************************************************************


From vic.cc.purdue.edu!abe@netcomsv.netcom.com Fri Sep  2 13:39:18 1994
To: ghoffman@ucsd.edu (Gary Hoffman)
Cc: freyr!pakers@netcomsv.uucp, birger@sdata.no, frode@ifi.unit.no,
        shadow@mindspring.com, pakers@netcom.uucp,
        hstoyan@faui80.informatik.uni-erlangen.de
Subject: Re: Do we need a GenWeb list? 
Date: Fri, 02 Sep 94 13:38:45 -0500

In message <1994Sep02.092942.316361@irpsbbs.ucsd.edu> you write:
>
>I am delighted with the recent discussions about HTML forms of genealogy
>databases. Before we get too far here, we need to consider whether the
>eight of us should form a new mailing list or whether we should continue
>these discussions in GEDCOM-L. 
>
>(I was recently bumped off of GEDCOM-L when my host was off line for a
>couple of hours. I just got back on. I therefore thank those who included
>me in their private traffic so that I could keep up during the GEDCOM-L
>blackout.) 
>
>After all, we are talking about genealogical data communication and GEDCOM
>itself. My vote is to keep these WWW Genealogy discussions in GEDCOM-L  and
>the rest of the world will follow. But if the majority of you believe
>otherwise, then let's start a new list and keep this GenWeb discussion
>moving.

I would like to continue this discussion the way it has begun,
rather than expanding it to GEDCOM-L.

>FYI: I have proposed a presentation on GenWeb at the WWW Conference in
>Chicago on October 17-19 and have invited Vic Abell to join me in the
>presentation. The notification of acceptance was delayed and I still don't
>know if they want this topic or not.  I am also considering a paper at
>GenTech in January, but I will be taking the California Bar Exam in
>February and I may need to study a bit for that ;-) Does anyone plan to
>attend GenTech and discuss GenWeb there?

I acknowledge Gary's gracious invitation, but haven't had time to
do anything about it.  I'm not sure I have enough by myself to
contribute.  Perhaps it would be better if Hary reported on the
work being done by all the members of this discussion group.

Vic

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Fri Sep  2 13:39:23 1994
To: Herbert Stoyan <hstoyan@faui80.informatik.uni-erlangen.de>
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: Re: gedcom-l and www 
Date: Fri, 02 Sep 94 13:45:23 -0500

In message <9409021707.AA25206@faui80.informatik.uni-erlangen.de> you write:
>
>I have nothing against gedcom-l. I did not look in it, yet. I used soc.roots.

I use soc.roots myself.  In general I prefer news groups to mailing lists.

>I'm interested how many www-people do you believe are working (like Abe)
>with Lynx? I'm removing all technical images just now out of the system.
>in one hour the system will be up again.

I'm not sure lynx is so much the issue as is the use of ASCII-only
interfaces of any kind -- e.g., people who use a modem to access a
system from which they can then access WWW usually have to use lynx.

BTW: I have both lynx and XMosaic available to me and try to use them
both when I'm viewing WWW things to see if they're suitable for both GUI
and ASCII readers.

>Separate issues are criteria of thrustworthness for data if we are willing to
>add a new tree in the net of trees. 1 year ago there was a guy who had a big
>mormon file of ascendants and descendants of Charlemagne. The descendants are
>fine, but the other part was nonsense. One should have something like a checke
>d
>kernel and an attachment procedure.

I tend to think that each GEDCOM data base should stand by itself with
a statement of ownership (and responsibility).  No one can avoid mistakes,
and the reader should beware.

Vic

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Fri Sep  2 13:39:25 1994
To: ghoffman@ucsd.edu (Gary Hoffman)
Cc: freyr!pakers@netcomsv.uucp, birger@sdata.no, frode@ifi.unit.no,
        shadow@mindspring.com, pakers@netcom.uucp,
        hstoyan@faui80.informatik.uni-erlangen.de
Subject: Re: Do we need a GenWeb list? 
Date: Fri, 02 Sep 94 13:38:45 -0500

In message <1994Sep02.092942.316361@irpsbbs.ucsd.edu> you write:
>
>I am delighted with the recent discussions about HTML forms of genealogy
>databases. Before we get too far here, we need to consider whether the
>eight of us should form a new mailing list or whether we should continue
>these discussions in GEDCOM-L. 
>
>(I was recently bumped off of GEDCOM-L when my host was off line for a
>couple of hours. I just got back on. I therefore thank those who included
>me in their private traffic so that I could keep up during the GEDCOM-L
>blackout.) 
>
>After all, we are talking about genealogical data communication and GEDCOM
>itself. My vote is to keep these WWW Genealogy discussions in GEDCOM-L  and
>the rest of the world will follow. But if the majority of you believe
>otherwise, then let's start a new list and keep this GenWeb discussion
>moving.

I would like to continue this discussion the way it has begun,
rather than expanding it to GEDCOM-L.

>FYI: I have proposed a presentation on GenWeb at the WWW Conference in
>Chicago on October 17-19 and have invited Vic Abell to join me in the
>presentation. The notification of acceptance was delayed and I still don't
>know if they want this topic or not.  I am also considering a paper at
>GenTech in January, but I will be taking the California Bar Exam in
>February and I may need to study a bit for that ;-) Does anyone plan to
>attend GenTech and discuss GenWeb there?

I acknowledge Gary's gracious invitation, but haven't had time to
do anything about it.  I'm not sure I have enough by myself to
contribute.  Perhaps it would be better if Hary reported on the
work being done by all the members of this discussion group.

Vic

From vic.cc.purdue.edu!abe@netcomsv.uucp Fri Sep  2 13:39:27 1994
To: Herbert Stoyan <hstoyan@faui80.informatik.uni-erlangen.de>
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: Re: GenWeb 
Date: Fri, 02 Sep 94 13:46:50 -0500

In message <9409021721.AA25251@faui80.informatik.uni-erlangen.de> you write:
>
>
>We should compile a list of genealogical urls.

That's a good idea, but I think a better one is to have a single
URL that has links to all the others.  Gary's WWW page is close,
I think.  How about it Gary, could you collect URLs and add them
to your page?

Vic

From freyr!pakers@netcomsv.netcom.com Sat Sep  3 12:36:50 1994
Subject: GenWeb
To: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        pakers@netcom.uucp
Date: Sat, 3 Sep 1994 12:26:37 -0700 (PDT)


Hello everyone...
First, I would like to say that I like the idea of this small sub-group
that Birger has brought together... we have a common interest in HTML
and geneology and I would rather have a chance to discuss this in a more
focused manner than in the GEDCOM-L list... which I am not a part of
anyway :-)

I will also say that I have almost no knowledge GEDCOM so I will have to
go collect the latest info and see what it is about...

It also occures to me that a person who has more than a couple hundred
people in his database has a completely different set of concerns than
what I have and is probably forced to be more impersonal with the data
than what I am trying to do... so it could be that we really are talking
about two different things... it is one thing to put together a HTML
presentation on what is basically one small family tree and another
thing entirely to have to worry about a 1000+ people database... I have
been looking at this as sort of an extended family album... a way to
organize all of the information that my parents have collected no matter
what format it is in... I would stop where we tie into larger family
groups that have already been researched... those larger family groups
would be the subject of other databases... where GEDCOM might come in
useful for me is in transferring information to other people in such a
way that it might be easly integrated into other databases... maybe
there could be something like the internet nameservers :-) where you
query a GEDCOM database for information and then insert it into your
own database as you see fit...

Phil
-- 

 Phillip L. Akers
 pakers@netcom.com
 freyr!pakers@netcomsv.netcom.com
 wa4dde@wa4dde.ampr.org (44.4.30.74)


From mindspring.com!shadow@netcomsv.netcom.com Sat Sep  3 15:37:13 1994
Date: Sat, 03 Sep 1994 18:32:10 -0500
To: freyr!pakers@netcomsv.uucp (Phillip Akers)
Subject: Re: GenWeb
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com


I guess that I'm sort of in the same boat as Phil when it comes to GEDCOM.
But I've subscribet to the GEDCOM-L list in hopes of learning more about it,
and the future development of it. 

I also think your right, Phil in saying that we may be talking about two
different things here. But I'm thinking that in the end result they will tie
in together quite nicely, given the possibility to link it in to others.
That possible linkage could be the key to bringing this all together. 

One of the things that I'm also thinking about is that what we might want to
do is put out some announcement posts on ROOTS-L, GEDCOM-L. soc.roots, and
alt.genealogy, as well as c.i.www.misc letting people know whats going on.
As I mentioned before, I know of several folks that have expressed to me a
interest in this, and I'm certain that they would love to be included in
this.  If no one has any objections to the idea, I'd be willing to do this.
My thinking is that the more input we can get here, the better off we are. 

Bill

**************************************************************************
 Bill Spurlock                               5125-C BeverlyGlen Lane
 shadow@mindspring.com      Norcross, GA  
 (404) 368-8884                          30092
**************************************************************************
SPURLOCK FAMILY GENEALOGY HOMEPAGE :
 http://www.mindspring.com/users/shadow/shadow.html
**************************************************************************


From faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com Mon Sep  5 01:54:54 1994
Date: Mon, 5 Sep 1994 10:46:49 --100
From: Herbert Stoyan <faui80.informatik.uni-erlangen.de!hstoyan@netcomsv.netcom.com>
To: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: GenWeb - local, global

The difference in the possible views is not the size. The difference is 
locality.
Local: If somebody is interested only to deliver his local web (generated from
one or more gedcom-files) into a set of local webs.
Global: If somebody is interested to link his local web into a global web and
to permit links to other subwebs being integrated in his local web.

My goal is the global GenWeb. I dream of a genealogical web which connects all
local webs into one big world wide genealogical web. We will start with sets of
local webs, that's clear. But soon, if two webbers detect common netparts,
they construct links.
For example, if Abe detects his line to Charlemagne, he knows that the last 10
generations are in my datebase in Erlangen. Instead of enlarging his gedcom-file
he incorporates a link to the person's url here in Erlangen.
If Birger finds out, some of his gggrandfathers is listed somewhere in Abes
local net. He incorporates a link.

It is impossible for everybody of us, even for a big organization like the
Mormons, to construct a database of all descendants of Charlemagne. As worldwide
distributed GenWeb this is possible.

GenWeb as set of local webs will be only a stage in the evolving web. Not for
long, and we have parts of the global web. The web will be interesting for new
addenda, because one will get new predecessors.

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Mon Sep  5 05:33:46 1994
To: freyr!pakers@netcomsv.uucp (Phillip Akers)
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        pakers@netcom.uucp
Subject: Re: GenWeb 
Date: Mon, 05 Sep 94 07:29:07 -0500

Phillip,

In message <9409031926.AA10977@freyr.uucp> you write:
>
>...
>
>I will also say that I have almost no knowledge GEDCOM so I will have to
>go collect the latest info and see what it is about...

I have put a copy of the latest GEDCOM specification (draft 5.3) on my
ftp server.  It's available via anonymous ftp from vic.cc.purdue.edu
(128.210.15.16) in pub/GEDCOM.std.gz.  It is a gzip'd text file.

>It also occures to me that a person who has more than a couple hundred
>people in his database has a completely different set of concerns than
>what I have and is probably forced to be more impersonal with the data
>than what I am trying to do... so it could be that we really are talking
>about two different things...

I don't think we have to have two different approaches.  The 5.3 GEDCOM
specification tries to address multi-media record forms, so it should be
possible to organize them via HTML.  What I was concerned with in my
first few attempts to organize some moderately sized GEDCOM data bases
with HTML was to make sure that the linkage process was efficient.  I
think that's a necessary goal whether the data is straight text, as mine
is, or whether it contains sight and sound.

What I would like to have a chnace to work with is a GEDCOM data base
that uses the new multi-media tags, so I could see what might be involved
in making it available via WWW.  Does anyone have such a thing or know of
a genealogy DB program that produces the tags?

Vic, W9RGB

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Mon Sep  5 05:33:48 1994
To: freyr!pakers@netcomsv.uucp (Phillip Akers)
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        pakers@netcom.uucp
Subject: Re: GenWeb 
Date: Mon, 05 Sep 94 07:29:07 -0500


Phillip,

In message <9409031926.AA10977@freyr.uucp> you write:
>
>...
>
>I will also say that I have almost no knowledge GEDCOM so I will have to
>go collect the latest info and see what it is about...

I have put a copy of the latest GEDCOM specification (draft 5.3) on my
ftp server.  It's available via anonymous ftp from vic.cc.purdue.edu
(128.210.15.16) in pub/GEDCOM.std.gz.  It is a gzip'd text file.

>It also occures to me that a person who has more than a couple hundred
>people in his database has a completely different set of concerns than
>what I have and is probably forced to be more impersonal with the data
>than what I am trying to do... so it could be that we really are talking
>about two different things...

I don't think we have to have two different approaches.  The 5.3 GEDCOM
specification tries to address multi-media record forms, so it should be
possible to organize them via HTML.  What I was concerned with in my
first few attempts to organize some moderately sized GEDCOM data bases
with HTML was to make sure that the linkage process was efficient.  I
think that's a necessary goal whether the data is straight text, as mine
is, or whether it contains sight and sound.

What I would like to have a chnace to work with is a GEDCOM data base
that uses the new multi-media tags, so I could see what might be involved
in making it available via WWW.  Does anyone have such a thing or know of
a genealogy DB program that produces the tags?

Vic, W9RGB

From mindspring.com!shadow@netcomsv.netcom.com Mon Sep  5 05:53:35 1994
Date: Mon, 05 Sep 1994 08:29:40 -0500
To: Herbert Stoyan <hstoyan@faui80.informatik.uni-erlangen.de>
From: mindspring.com!shadow@netcomsv.netcom.com (Bill Spurlock)
Subject: Re: GenWeb - local, global
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        shadow@mindspring.com, pakers@netcom.uucp


>My goal is the global GenWeb. I dream of a genealogical web which connects all
>local webs into one big world wide genealogical web. We will start with sets of
>local webs, that's clear. But soon, if two webbers detect common netparts,
>they construct links.

This is exactly my thinking here as well. The potential here is beyond
anything else that we have today. 

The key to making this work is, of course, the encouragement and development
of the local webs. No doubt it will take some time to get such a network in
place, but it can happen. 

I'm also thinking that a large network of locals will add to the integrity,
creating a built in check. Say for example, my line connects with that of
Phil at some point. The data that Phil has that overlaps contains several
major differences. After talking it over and looking at both sets of data,
we find that Phils data is "stronger" than mine, for whatever reasons (tho I
would suspect source verification). So his data is linked to mine, we and we
have corrected a problem ( as well as future problems) of passing along bad
or incorrect information. In effect, what we are doing here is making
genealogy a great example of teamwork. And we can sidestep to some extent
the bad data problem that exists in the LDS and other places. Mind you, we
will most likely never eliminate the problem, but it is a small step in the
right direction 

Bill


**************************************************************************
 Bill Spurlock                               5125-C BeverlyGlen Lane
 shadow@mindspring.com      Norcross, GA  
 (404) 368-8884                          30092
**************************************************************************
SPURLOCK FAMILY GENEALOGY HOMEPAGE :
 http://www.mindspring.com/users/shadow/shadow.html
**************************************************************************


From vic.cc.purdue.edu!abe@netcomsv.netcom.com Mon Sep  5 05:53:36 1994
To: Herbert Stoyan <hstoyan@faui80.informatik.uni-erlangen.de>
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: Re: GenWeb - local, global 
Date: Mon, 05 Sep 94 07:40:33 -0500

Herbert,

In message <9409050846.AA03852@faui80.informatik.uni-erlangen.de> you write:
>
>
>The difference in the possible views is not the size. The difference is 
>locality.
>Local: If somebody is interested only to deliver his local web (generated from
>one or more gedcom-files) into a set of local webs.
>Global: If somebody is interested to link his local web into a global web and
>to permit links to other subwebs being integrated in his local web.

Hmm.  You've touched on an issue we haven't discussed yet -- privacy.
I discovered this when starting up my experimental GEDCOM->WWW
service.  Someone provided me a data base; I made it available on
my server; later he asked me to remove it, because he was uncomfortable
having some particularly painful aspects of his family history
(holocaust records) so widely available.

Having offered that diversion, I'm not sure I see where your locality
argument leads.  Once a genealogy data base becomes available via
WWW, I don't think it's possible to restrict access to a "local"
domain.  If it's readable, anyone can build links to it.

What interests me more is how a large, dispersed set of data bases
can be linked together.  For example, I have a server with 5 data
bases.  Assuming that each of you has a similar server, will there
be a master index someplace?  How will that be done?  Who will
maintain it?

>My goal is the global GenWeb. I dream of a genealogical web which connects all
>local webs into one big world wide genealogical web. We will start with sets o
>f
>local webs, that's clear. But soon, if two webbers detect common netparts,
>they construct links.
>For example, if Abe detects his line to Charlemagne, he knows that the last 10

Alert!  "Abe" is one of my least favorite nicknames.  "Vic" is just
as short and easy to type.  I would prefer that it be used instead.

>generations are in my datebase in Erlangen. Instead of enlarging his gedcom-fi
>le
>he incorporates a link to the person's url here in Erlangen.
>If Birger finds out, some of his gggrandfathers is listed somewhere in Abes
>local net. He incorporates a link.

This is the nature of WWW, and an important reason why it is a natural
vehicle for genealogical information interchange.  I hope we can conceive
a reasonable, practical, and efficient way that dispersed data can be
interconnected.

Vic

From mindspring.com!shadow@netcomsv.netcom.com Mon Sep  5 06:55:01 1994
Date: Mon, 05 Sep 1994 09:27:06 -0500
To: Herbert Stoyan <hstoyan@faui80.informatik.uni-erlangen.de>
Subject: GenWeb - Importing Methods
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        shadow@mindspring.com, pakers@netcom.uucp
X-Mailer: <Windows Eudora Version 1.4.2b16>


After thinking about the possible ways of importing data to a HTML formatted
page, I'm wondering if it might be wise to look at the possibility of using
a alternate method as opposed to gedcom. 

What I'm thinking of is,taking paf as a example, creating a utility program
that would A) allow for the selection of individual record numbers to
include in the process, and B), create a set of pre formatted HTLM pages
with all links in place, with the obivious exception of the links to other
data sources..that is to say all internal links would be set by the utility.

I'm also thinking that A HTML to GEDCOM converter might be a interesting
addition for folks that would like to export data into their database from
the HTML doccuments.

Suggestions? Comments?

Bill

**************************************************************************
 Bill Spurlock                               5125-C BeverlyGlen Lane
 shadow@mindspring.com      Norcross, GA  
 (404) 368-8884                          30092
**************************************************************************
SPURLOCK FAMILY GENEALOGY HOMEPAGE :
 http://www.mindspring.com/users/shadow/shadow.html
**************************************************************************


From freyr!pakers@netcomsv.netcom.com Mon Sep  5 10:53:20 1994
Subject: GenWeb... Local & Global
To: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        pakers@netcom.uucp
Date: Mon, 5 Sep 1994 10:05:06 -0700 (PDT)

Hmmm... I think the way to do this is to follow the example given to
us by WWW... after all, what is it but a gigantic distributed database
of whatever anyone wants to put on the Web... for instance, the way to
handle the indexing is by breaking it up... just as it is done on the
Web... you do not want to have to load one file that has 9000 people in
it for an index... instead the top-level file might only contain surnames
and when a surname is selected then that takes you to another file that
contains a list of people with that surname... an example of this can be
seen in my surname list... it is also the way that most indexes are built
on the Web... someone will do a meta-index which may call an index on a
specific subject, which in turn may call a local index on a very specific
part of that subject... you just walk your way through till you find what
you want... with forms capability then it is possible to maybe even add
short cuts... 

It really is a matter of how one choses to organize the data... by going
with smaller files one does add more files but also speeds up the process...
so like the Web maybe we have to spread the files out over more systems to
avoid overloading any one system...

As for linking the databases, they just link together at the shared names
and should be transparent to the user... they may not even know they have 
moved from a database in the US to one in Norway or Germany... that is
the beauty of the system... it is not even necessary for the databases
to have the same format... someone may be browsing Vics GEDCOM database
and follow a link over to my HTML database by selecting a name... 

Phil
-- 

 Phillip L. Akers
 pakers@netcom.com
 freyr!pakers@netcomsv.netcom.com
 wa4dde@wa4dde.ampr.org (44.4.30.74)
 

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Tue Sep  6 05:34:11 1994
To: shadow@mindspring.com (Bill Spurlock)
Cc: Herbert Stoyan <hstoyan@faui80.informatik.uni-erlangen.de>,
        birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        pakers@netcom.uucp
Subject: Re: GenWeb - Importing Methods 
Date: Tue, 06 Sep 94 07:30:06 -0500

In message <199409051324.JAA02646@dylan.mindspring.com> you write:
>
>
>After thinking about the possible ways of importing data to a HTML formatted
>page, I'm wondering if it might be wise to look at the possibility of using
>a alternate method as opposed to gedcom. 
>
>What I'm thinking of is,taking paf as a example, creating a utility program
>that would A) allow for the selection of individual record numbers to
>include in the process, and B), create a set of pre formatted HTLM pages
>with all links in place, with the obivious exception of the links to other
>data sources..that is to say all internal links would be set by the utility.

But now you're created a new "standard."  GEDCOM is supposed to be the
interchange standard, and I favor sticking with it.

>I'm also thinking that A HTML to GEDCOM converter might be a interesting
>addition for folks that would like to export data into their database from
>the HTML doccuments.

This is an intersting idea, worth pursuing in my opinion.

Vic

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Tue Sep  6 05:53:54 1994
To: freyr!pakers@netcomsv.uucp (Phillip Akers)
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        pakers@netcom.uucp
Subject: Re: GenWeb... Local & Global 
Date: Tue, 06 Sep 94 07:50:37 -0500

In message <9409051705.AA15103@freyr.uucp> you write:
>
>Hmmm... I think the way to do this is to follow the example given to
>us by WWW... after all, what is it but a gigantic distributed database
>of whatever anyone wants to put on the Web... for instance, the way to
>handle the indexing is by breaking it up... just as it is done on the
>Web... you do not want to have to load one file that has 9000 people in
>it for an index... instead the top-level file might only contain surnames
>and when a surname is selected then that takes you to another file that
>contains a list of people with that surname... an example of this can be

Perhaps you need to look at what I have done with that 9,000+ person
data base.  (The URL is "http://www-rcd.cc.purdue.edu/abe/".)  The
large Bratt data base has 18 name-range URLs at the top level.
Each name range URL points at a separate index page of about 40KB.
(I chose that size with an eye on comfortable net transmission and
a reasonable link count that viewers could handle effectively.)
BTW: the division of the GEDCOM source into these segments is done
automatically with a single, simple C program.

>seen in my surname list... it is also the way that most indexes are built
>on the Web... someone will do a meta-index which may call an index on a
>specific subject, which in turn may call a local index on a very specific
>part of that subject... you just walk your way through till you find what
>you want... with forms capability then it is possible to maybe even add
>short cuts... 

>It really is a matter of how one choses to organize the data... by going
>with smaller files one does add more files but also speeds up the process...
>so like the Web maybe we have to spread the files out over more systems to
>avoid overloading any one system...

I don't think large numbers of smaller files is the answer.  Even
a data base like mine creates too many files for one system.  And
I don't think we can afford one system per data base.

We must face the fact that HTML links are inefficient when used in
large numbers.  That's why the CGI (Common Gateway Interface)
specification exists -- to allow efficient treatments of large link
counts in ways that are far more efficient than URLs.

>As for linking the databases, they just link together at the shared names
>and should be transparent to the user... they may not even know they have 
>moved from a database in the US to one in Norway or Germany... that is
>the beauty of the system... it is not even necessary for the databases
>to have the same format... someone may be browsing Vics GEDCOM database
>and follow a link over to my HTML database by selecting a name... 

I'm not sure I see how the links will be maintained.  Embedding
them directly in individual data bases is too rigid from my
experience.  After all, host names change, file locations change,
URLs are fairly complicated and subject to change themselves.  What
person among us hasn't tried to use a URL and found a typo in it?
Or returned to a favorite URL to find that a component has changed?
I suspect the dispersion will have to be much more complicated.
Perhaps the DNS is a good model with surname servers that translate
meta-links into URLs, just as name servers translate names into
network numbers.

Vic

From vest.sdata.no!Birger.Wathne@netcomsv.netcom.com Tue Sep  6 07:34:43 1994
Date: Tue, 6 Sep 94 15:39:14 +0200
To: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no, abe@cc.purdue.edu,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        pakers@netcom.uucp, freyr!pakers@netcomsv.uucp
Subject: Re: GenWeb... Local & Global


>
>As for linking the databases, they just link together at the shared names
>and should be transparent to the user... they may not even know they have 
>moved from a database in the US to one in Norway or Germany... that is
>the beauty of the system... it is not even necessary for the databases
>to have the same format... someone may be browsing Vics GEDCOM database
>and follow a link over to my HTML database by selecting a name... 
>
>Phil
>-- 
>
> Phillip L. Akers
> pakers@netcom.com
> freyr!pakers@netcomsv.netcom.com
> wa4dde@wa4dde.ampr.org (44.4.30.74)
> 

But how should this be expressed in the GEDCOM files for those of us
who want to use GEDCOM as the main database format? Should we set up
an empty dummy person with a special tag that says 'this person is
available from another WWW server', and enough information to
look the person up automatically?

Or should we somehow encode it directly in the reference to child/spouse?

As I agree that just hard-coding URL's into the databases is risky, we should
have some kind of lookup mechanism. I think I would prefer the first approach.
I guess this is something that will have to be decided on as we get something
working. But whatever we decide on should be flexible, as we'll certainly
want to change it sometime.


Birger

From vest.sdata.no!Birger.Wathne@netcomsv.netcom.com Tue Sep  6 07:34:45 1994
Date: Tue, 6 Sep 94 15:31:50 +0200
To: freyr!pakers@netcomsv.uucp, abe@cc.purdue.edu
Subject: Re: GenWeb
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        hstoyan@faui80.informatik.uni-erlangen.de, shadow@mindspring.com,
        pakers@netcom.uucp

>
>What I would like to have a chnace to work with is a GEDCOM data base
>that uses the new multi-media tags, so I could see what might be involved
>in making it available via WWW.  Does anyone have such a thing or know of
>a genealogy DB program that produces the tags?
>
>Vic, W9RGB
>

Well, Lifelines lets you work directly with any tag you want to use.
It just gives you a default set of tags to fill out on any new entity,
but you can then add GEDCOM tags as you wish. So the multimedia tag
will be available. Of course lifelines won't do anything sensible with
the data yet. You can set up a macro in your favourite editor to
spawn xv, mpeg_play or audiotool from the tagged-line, of course.

This tag should just contain a reference to a file. The GED2HTML
filter should then either include this as part of the page
(for certain special tags, perhaps?), or create a link to the
picture/movie/sound.


Birger

From vest.sdata.no!Birger.Wathne@netcomsv.netcom.com Tue Sep  6 07:34:47 1994
Date: Tue, 6 Sep 94 15:34:15 +0200
To: hstoyan@faui80.informatik.uni-erlangen.de, abe@cc.purdue.edu
Subject: Re: GenWeb - local, global
Cc: birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        shadow@mindspring.com, pakers@netcom.uucp


>
>Hmm.  You've touched on an issue we haven't discussed yet -- privacy.
>I discovered this when starting up my experimental GEDCOM->WWW
>service.  Someone provided me a data base; I made it available on
>my server; later he asked me to remove it, because he was uncomfortable
>having some particularly painful aspects of his family history
>(holocaust records) so widely available.
>
>
>Vic
>

Each GEDCOM tag can have several levels of information. Perhaps we
could use a subtag to indicate that this tag is not meant for public
access, and then fix the GED2HTML filter so these tags don't get
included?


Birger

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Tue Sep  6 08:54:16 1994
To: Birger.Wathne@vest.sdata.no (Birger A. Wathne)
Cc: freyr!pakers@netcomsv.uucp, birger@sdata.no, ghoffman@ucsd.edu,
        frode@ifi.unit.no, hstoyan@faui80.informatik.uni-erlangen.de,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: Re: GenWeb 
Date: Tue, 06 Sep 94 10:40:01 -0500

In message <9409061331.AA05958@sdvest> you write:
>
>Well, Lifelines lets you work directly with any tag you want to use.
>It just gives you a default set of tags to fill out on any new entity,
>but you can then add GEDCOM tags as you wish. So the multimedia tag
>will be available. Of course lifelines won't do anything sensible with
>the data yet. You can set up a macro in your favourite editor to
>spawn xv, mpeg_play or audiotool from the tagged-line, of course.

I would like to see a GEDCOM file where these tags are really used.

>This tag should just contain a reference to a file. The GED2HTML
>filter should then either include this as part of the page
>(for certain special tags, perhaps?), or create a link to the
>picture/movie/sound.

Yes, I think this part of the process is pretty clear.  The eventual
DB will need an entry that a CGI processor, for example, can convert
into a URL for the file.

Vic

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Tue Sep  6 08:54:17 1994
To: Birger.Wathne@vest.sdata.no (Birger A. Wathne)
Cc: freyr!pakers@netcomsv.uucp, birger@sdata.no, ghoffman@ucsd.edu,
        frode@ifi.unit.no, hstoyan@faui80.informatik.uni-erlangen.de,
        shadow@mindspring.com, pakers@netcom.uucp
Subject: Re: GenWeb 
Date: Tue, 06 Sep 94 10:40:01 -0500


In message <9409061331.AA05958@sdvest> you write:
>
>Well, Lifelines lets you work directly with any tag you want to use.
>It just gives you a default set of tags to fill out on any new entity,
>but you can then add GEDCOM tags as you wish. So the multimedia tag
>will be available. Of course lifelines won't do anything sensible with
>the data yet. You can set up a macro in your favourite editor to
>spawn xv, mpeg_play or audiotool from the tagged-line, of course.

I would like to see a GEDCOM file where these tags are really used.

>This tag should just contain a reference to a file. The GED2HTML
>filter should then either include this as part of the page
>(for certain special tags, perhaps?), or create a link to the
>picture/movie/sound.

Yes, I think this part of the process is pretty clear.  The eventual
DB will need an entry that a CGI processor, for example, can convert
into a URL for the file.

Vic

From vic.cc.purdue.edu!abe@netcomsv.netcom.com Tue Sep  6 08:54:19 1994
To: Birger.Wathne@vest.sdata.no (Birger A. Wathne)
Cc: hstoyan@faui80.informatik.uni-erlangen.de, birger@sdata.no,
        ghoffman@ucsd.edu, frode@ifi.unit.no, shadow@mindspring.com,
        pakers@netcom.uucp
Subject: Re: GenWeb - local, global 
Date: Tue, 06 Sep 94 10:43:29 -0500

In message <9409061334.AA05963@sdvest> you write:
>
>>Hmm.  You've touched on an issue we haven't discussed yet -- privacy.
>>I discovered this when starting up my experimental GEDCOM->WWW
>>service.  Someone provided me a data base; I made it available on
>>my server; later he asked me to remove it, because he was uncomfortable
>>having some particularly painful aspects of his family history
>>(holocaust records) so widely available.
>
>Each GEDCOM tag can have several levels of information. Perhaps we
>could use a subtag to indicate that this tag is not meant for public
>access, and then fix the GED2HTML filter so these tags don't get
>included?

I don't know what would suffice.  When I questioned the person who
wanted me to cancel his GEDCOM/WWW submission, he never answered
my detailed questions about what was objectionable and what he was
willing to share.  In his case, I can guess that he was uncomfortable
sharing information about deaths and notes of a personal nature
related to the Holocaust, but he never became explicit enough to
me that I could have decided which GEDCOM tags to ignore.

You may have the right idea, however.

Vic

From mindspring.com!shadow@netcomsv.netcom.com Tue Sep  6 14:03:48 1994
Date: Tue, 06 Sep 1994 12:08:22 -0500
To: abe@cc.purdue.edu
Subject: Re: GenWeb - local, global 
Cc: hstoyan@faui80.informatik.uni-erlangen.de, birger@sdata.no,
        ghoffman@ucsd.edu, frode@ifi.unit.no, shadow@mindspring.com,
        pakers@netcom.uucp

>In message <9409061334.AA05963@sdvest> you write:
>>
>>>Hmm.  You've touched on an issue we haven't discussed yet -- privacy.
>>>I discovered this when starting up my experimental GEDCOM->WWW
>>>service.  Someone provided me a data base; I made it available on
>>>my server; later he asked me to remove it, because he was uncomfortable
>>>having some particularly painful aspects of his family history
>>>(holocaust records) so widely available.
>>
>>Each GEDCOM tag can have several levels of information. Perhaps we
>>could use a subtag to indicate that this tag is not meant for public
>>access, and then fix the GED2HTML filter so these tags don't get
>>included?
>
>I don't know what would suffice.  When I questioned the person who
>wanted me to cancel his GEDCOM/WWW submission, he never answered
>my detailed questions about what was objectionable and what he was
>willing to share.  In his case, I can guess that he was uncomfortable
>sharing information about deaths and notes of a personal nature
>related to the Holocaust, but he never became explicit enough to
>me that I could have decided which GEDCOM tags to ignore.

I have to disagree with this whole concept here. Please correct me if I'm
wrong , but is not the purpose of GEDCOM to allow a person to export their
data to a different data source so that it may be used by other?  If that
person has information that he or she feels should be withheld for whatever
reason, then ( and I do NOT agree with this idea or support it any way, I'm
just offering a point here) that information should not be included in the
GEDOCM.

Most of us have run into this sort of a problem in our research at some
point, someone covering up vital information because they feel that they are
protecting the family in some manner. This is wrong, and should not be
incouraged IMHO . 

Bill
**************************************************************************
 Bill Spurlock                               5125-C BeverlyGlen Lane
 shadow@mindspring.com      Norcross, GA  
 (404) 368-8884                          30092
**************************************************************************
SPURLOCK FAMILY GENEALOGY HOMEPAGE :
 http://www.mindspring.com/users/shadow/shadow.html
**************************************************************************


From vic.cc.purdue.edu!abe@netcomsv.netcom.com Tue Sep  6 14:03:57 1994
To: Birger.Wathne@vest.sdata.no (Birger A. Wathne)
Cc: shadow@mindspring.com, hstoyan@faui80.informatik.uni-erlangen.de,
        birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        pakers@netcom.uucp
Subject: Re: GenWeb - local, global 
Date: Tue, 06 Sep 94 12:47:50 -0500

In message <9409061640.AA06344@sdvest> you write:
>
>>>So let us keep privacy out of the GEDCOM and HTML discussions.
>>>It should be done when exporting data to GEDCOM in the first place.
>>
>>While I agree that editing data before it's introduced into the
>>WWW is the best alternative, I suspect that some submitters will
>>not realize how widely their information is being dispersed.  Some
>>will not have the technical savvy to do the editing.  Some may
>>prefer that it be performed centrally.  The GEDCOM to WWWW interface
>>might be a natural place to perform automatic editing based on
>>tags.
>>
>But how could the GEDCOM to HTML interface know which tags to exclude
>unless the provider had tagged it? And if he can tag it, he should
>be able to remove it as well?

No, not at all.  Many people take a black box approach to genealogical
computing.  Their genealogical DB program might produce tags that
they neither understand nor recognize.  They probably can state
that death places (or birth locations, or whatever) should be
excluded without needing to understand how GEDCOM tags the information
and whether or not their DB can be told to exclude them when
producing GEDCOM.

>I agree that submitters should be warned. Perhaps someone could set up
>a standard GenWEB submitter agreement?

Sounds like a reasonable appoach.  Do we have to involve the lawyers?  :-)

Vic

From vest.sdata.no!Birger.Wathne@netcomsv.netcom.com Tue Sep  6 14:03:59 1994
Date: Tue, 6 Sep 94 18:11:24 +0200
To: abe@cc.purdue.edu, shadow@mindspring.com
Subject: Re: GenWeb - local, global
Cc: hstoyan@faui80.informatik.uni-erlangen.de, birger@sdata.no,
        ghoffman@ucsd.edu, frode@ifi.unit.no, pakers@netcom.uucp

>
>I have to disagree with this whole concept here. Please correct me if I'm
>wrong , but is not the purpose of GEDCOM to allow a person to export their
>data to a different data source so that it may be used by other?  If that
>person has information that he or she feels should be withheld for whatever
>reason, then ( and I do NOT agree with this idea or support it any way, I'm
>just offering a point here) that information should not be included in the
>GEDOCM.
>
>Most of us have run into this sort of a problem in our research at some
>point, someone covering up vital information because they feel that they are
>protecting the family in some manner. This is wrong, and should not be
>incouraged IMHO . 
>
>Bill

I can agree that perhaps we shouldn't concern ourselves with this.
It's up to the data provider to ensure that the gedcom data used as
source for the ged2html conversion doesn't contain anything they
don't want published.

Someone offering GEDCOM files for publishing should edit out everything
they want to keep private.

Someone extracting data through a CGI gateway (like me) could use
the CGI program to check who's reading the data, and filter out
'private' stuff in the report generating the GEDCOM file.

So let us keep privacy out of the GEDCOM and HTML discussions.
It should be done when exporting data to GEDCOM in the first place.


Birger

From mindspring.com!shadow@netcomsv.netcom.com Tue Sep  6 14:04:01 1994
Date: Tue, 06 Sep 1994 13:03:44 -0500
To: abe@cc.purdue.edu
Subject: Re: GenWeb - local, global 
Cc: shadow@mindspring.com, hstoyan@faui80.informatik.uni-erlangen.de,
        birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        pakers@netcom.uucp

>While I agree that editing data before it's introduced into the
>WWW is the best alternative, I suspect that some submitters will
>not realize how widely their information is being dispersed.  Some
>will not have the technical savvy to do the editing.  Some may
>prefer that it be performed centrally.  The GEDCOM to WWWW interface
>might be a natural place to perform automatic editing based on
>tags.
>
>Invitations to submit information for WWW access at least need a
>warning or disclaimer about the public nature of the service.  I
>don't think we can keep silent about this aspect.  I did and it
>crept up on me by surprise.

While I would strongly agree with the idea of a warning or disclaimer in
regards to public access, the idea of a automatic editing servics I could
not agree with.  As I see it, most people have a good idea of what a GEDCOM
is before they create it, and realize that it will be seen by any number of
people, no matter what the distrubution method is,  It would be only fair on
our part to ensure that a potential partipicant has a understanding of what
he or she is getting involved in before the submission, and it it up to them
to submit data that is acceptable to themselves. But creating a service that
supports the concept of covering up data because the author feels that it
should not be publish is not right. Again, if the author feels that the data
is "sensitive" then they have the option of leaving that part out of the
submission.  But that is something that we should not be involved in in any
manner save the warning.

Bill
**************************************************************************
 Bill Spurlock                               5125-C BeverlyGlen Lane
 shadow@mindspring.com      Norcross, GA  
 (404) 368-8884                          30092
**************************************************************************
SPURLOCK FAMILY GENEALOGY HOMEPAGE :
 http://www.mindspring.com/users/shadow/shadow.html
**************************************************************************


From vic.cc.purdue.edu!abe@netcomsv.netcom.com Tue Sep  6 14:04:02 1994
To: Birger.Wathne@vest.sdata.no (Birger A. Wathne)
Cc: shadow@mindspring.com, hstoyan@faui80.informatik.uni-erlangen.de,
        birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        pakers@netcom.uucp
Subject: Re: GenWeb - local, global 
Date: Tue, 06 Sep 94 11:41:13 -0500

In message <9409061611.AA06305@sdvest> you write:
>
>Someone extracting data through a CGI gateway (like me) could use
>the CGI program to check who's reading the data, and filter out
>'private' stuff in the report generating the GEDCOM file.

Interesting.  What do you know other than the host that originates
the request?  Do you have any more personal information on the
identity of the source?

>So let us keep privacy out of the GEDCOM and HTML discussions.
>It should be done when exporting data to GEDCOM in the first place.

While I agree that editing data before it's introduced into the
WWW is the best alternative, I suspect that some submitters will
not realize how widely their information is being dispersed.  Some
will not have the technical savvy to do the editing.  Some may
prefer that it be performed centrally.  The GEDCOM to WWWW interface
might be a natural place to perform automatic editing based on
tags.

Invitations to submit information for WWW access at least need a
warning or disclaimer about the public nature of the service.  I
don't think we can keep silent about this aspect.  I did and it
crept up on me by surprise.

Vic

From vest.sdata.no!Birger.Wathne@netcomsv.netcom.com Tue Sep  6 14:04:06 1994
Date: Tue, 6 Sep 94 18:40:43 +0200
To: Birger.Wathne@vest.sdata.no, abe@cc.purdue.edu
Subject: Re: GenWeb - local, global
Cc: shadow@mindspring.com, hstoyan@faui80.informatik.uni-erlangen.de,
        birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        pakers@netcom.uucp


>>
>>Someone extracting data through a CGI gateway (like me) could use
>>the CGI program to check who's reading the data, and filter out
>>'private' stuff in the report generating the GEDCOM file.
>
>Interesting.  What do you know other than the host that originates
>the request?  Do you have any more personal information on the
>identity of the source?
>

Maybe not, but 


>>So let us keep privacy out of the GEDCOM and HTML discussions.
>>It should be done when exporting data to GEDCOM in the first place.
>
>While I agree that editing data before it's introduced into the
>WWW is the best alternative, I suspect that some submitters will
>not realize how widely their information is being dispersed.  Some
>will not have the technical savvy to do the editing.  Some may
>prefer that it be performed centrally.  The GEDCOM to WWWW interface
>might be a natural place to perform automatic editing based on
>tags.
>
But how could the GEDCOM to HTML interface know which tags to exclude
unless the provider had tagged it? And if he can tag it, he should
be able to remove it as well?

>Invitations to submit information for WWW access at least need a
>warning or disclaimer about the public nature of the service.  I
>don't think we can keep silent about this aspect.  I did and it
>crept up on me by surprise.
>
>Vic
>

I agree that submitters should be warned. Perhaps someone could set up
a standard GenWEB submitter agreement?


Birger

From vest.sdata.no!Birger.Wathne@netcomsv.netcom.com Tue Sep  6 14:04:09 1994
Date: Tue, 6 Sep 94 18:53:33 +0200
To: Birger.Wathne@vest.sdata.no, abe@cc.purdue.edu
Subject: Re: GenWeb - local, global
Cc: shadow@mindspring.com, hstoyan@faui80.informatik.uni-erlangen.de,
        birger@sdata.no, ghoffman@ucsd.edu, frode@ifi.unit.no,
        pakers@netcom.uucp


>>>
>>>Someone extracting data through a CGI gateway (like me) could use
>>>the CGI program to check who's reading the data, and filter out
>>>'private' stuff in the report generating the GEDCOM file.
>>
>>Interesting.  What do you know other than the host that originates
>>the request?  Do you have any more personal information on the
>>identity of the source?
>>
>
>Maybe not, but 


I guess this didn't tell anyone much :^)

Ok. Here is what I meant to say:

CGI gateways get info about the client's hostname and IP address.
It's also possible for the CGI script to demand some kind of user
authentication, but I don't know much about it yet.

In most cases, this could prove to be enough security, I guess.


Birger

From ucsd.edu!list-relay@netcomsv.netcom.com Fri Sep  9 14:33:31 1994
To: genweb@ucsd.edu
Date: Fri, 09 Sep 1994 14:18:45  PDT
Subject: GenWeb 2000

Welcome to the genweb mailing list. 

After I launched my proposal that came to be called GenWeb, my poor mailer
could not keep up with all e-mail traffic. So I got the UCSD mail managers
to give us a list on their listserver. (Maybe I'll do this for Christmas
Cards, too, this year.)

The prospect of creating a world-wide web formatted database of on-line
genealogy resources has fired the imagination of many people who have large
data archives to contribute. My proposal was widely circulated in Roots-L
and other electronic fora and recently has generated much activity. To get
up to speed, user Mosaic or another WWW browser to read my proposal at URL
http://irpsbbs.ucsd.edu/gene/genedemo.html.

The following is an excerpt of an article of mine to be published in the
September issue of COMPU.GEN, the newletter of the Computer Genealogy
Society of San Diego.  If you follow the line of reasoning, please give me
your reaction: is GenWeb 2000 achievable?
---------------------
[follows an introduction of WWW and GenWeb]
Indexing and Searching GenWeb Archives. With information spread among
several databases on widely-separated hosts, there are two approaches to
locating information on a particular person: indexing or searching. Using
the searching method, you could run a program on your computer that queries
each of the known GenWeb archives and reports to you the URL of individuals
that it finds that meet your search criteria. Such a search program would
build a special HTML page out of its search and then let you click on the
links that take you to various GenWeb archives to see if they hold
information on the person you are seeking. A variation on this program is
known as an agent, which is a program similar to a network virus that roams
the network in search of your ancestor and then sends you e-mail when it
finds someone you are searching for. This type of program is also called a
worm and could live in the network constantly checking GenWeb archives for
years until it expires or fulfills its mission. Neither of these search
programs exist yet, but could easily be made using current programming
techniques.

The other approach to finding an ancestor in GenWeb is to consult a master
index to all GenWeb archives. Constructing and maintaining such a master
index would involve the dedication of a large hard drive at the index site
plus the effort to keep the index updated by running some variation of the
search programs mentioned above. Presently, a World Wide Web Worm checks
each of the 4.2 billion possible Internet addresses to see if a WWW server
answers. If so, it registers that server's default or home page on its
index. Under the GenWeb concept, there would be fewer servers but many more
than one document per server to search. I envision indexes being used in 
the early years of GenWeb archive building, but eventually, I believe that
agent searches will end up being more helpful to people using GenWeb for
peidgree searches. 

Using GenWeb. Under GenWeb, the typical user will maintain information on
his or her immediate family inside their own computer. Writing HTML
documenets will become common, especially as word processor macros are
developed to simplify the task. Mosaic and other browsers can read local
files as well as networked documents, so it becomes a simple matter to
point Mosaic to the local files for information on near relatives. Most
people would keep track of their children, their siblings, and their
parents and their parents' siblings on their local mass storage device.
Beyond about the third of fourth generation, however, it would be more
efficient to allow genealogical information to be stored in a GenWeb
archive and link the local files to the archived files via a GenWeb
hypertext link. 

Building GenWeb Archives. Under this vision of use, the problems of privacy
issues that plague large databases as the LDS Church's Ancestral File are
avoided. Information on living persons is maintained locally in their or a
close relative's own computer. But back three or four generations,
information on people who are deceased would be maintained in GenWeb
archives which anyone could access. I estimate that the GenWeb concept
would be most helpful for individuals born before about 1910 or 75 years
ago. In fact, I am proposing that a major effort be launched to convert all
existing databases to GenWeb archives for persons born before the current
century and that the target complettion date for this project be set at the
end of the century, only 5 years away. Under the watchword "GenWeb 2000,"
computer genealogists who maintain any type of computer records would be
encouraged to submit those records to a GenWeb archivist who would then
make them available over the Internet. Individuals or institutions could
maintain such archives under common agreement with all other GenWeb sites.
In fact, the Computer Genealogy Society of San Diego could operate and
maintain a GenWeb archive as a public service to the genealogy community. I
believe the GenWeb 2000 goal is easily attainable if the idea were
promulgated widely.


***************************************************************************
*Gary B. Hoffman, Computer/Language Lab Director e-mail: ghoffman@ucsd.edu*
*Graduate School of International Relations and Pacific Studies    (IR/PS)*
*University of California, San Diego (UCSD)          voice: (619) 534-7733*
*9500 Gilman Dr., La Jolla, CA 92093-0519 USA          fax: (619) 534-3939*
***************************************************************************


From ucsd.edu!ghoffman@netcomsv.netcom.com Fri Sep  9 17:53:28 1994
Date: Fri, 09 Sep 1994 16:51:28  PDT
Subject: Announcing GenWeb Mailing List


If you are reading this message, you once corresponded with me regarding my
proposal to use the World Wide Web to distribute genealogy information. I
had originally called this CyberRoots, but at Jon Brinkmann's suggestion,
it is now dubbed "GenWeb".

This message announces a new Internet mailing list to support discussions
among those interested in implementing a genealogical information exchange
system using the World Wide Web, dubbed "GenWeb". 

The mailing list is called 'genweb' and is hosted at the University of
California at San Diego. To subscribe, please send a message to'
listserv@ucsd.edu' with the following text in the body of the message: 
   add genweb
To post a message to all subscribers of the list, address it to
'genweb@ucsd.edu'.

We hope you will join us in this next phase of computer accelerated
genealogy.

Gary


***************************************************************************
*Gary B. Hoffman, Computer/Language Lab Director e-mail: ghoffman@ucsd.edu*
*Graduate School of International Relations and Pacific Studies    (IR/PS)*
*University of California, San Diego (UCSD)          voice: (619) 534-7733*
*9500 Gilman Dr., La Jolla, CA 92093-0519 USA          fax: (619) 534-3939*
***************************************************************************


From UCSD.EDU!list-relay@netcomsv.netcom.com Sat Sep 10 17:33:27 1994
To: genweb@UCSD.EDU
Subject: GenWeb2000:  
Date: Sat, 10 Sep 1994 17:18:58
Cc: svpafug@rahul.net

TO: GenWeb Mailing List
FROM:  Bill Minnick, VP,
               Silicon Valley PAF User's Group

Many of our 1500 SVPAFUG  members are interested in doing what we can to 
support Gary Hoffman's GenWeb 2000 proposal.  We have a team of highly skilled 
professionals who also have a deep interest in computer genealogy.   Learn 
more about SVPAFUG by exploring our Home Page at URL:

http://www.rahul.net/svpafug

Please also look over the "Alternate Demo" we have place in our Home Page.

We want to participate in the design and building GenWeb.  We hope this 
group will consider the following in laying out the design of GenWeb:

1.  ease of data submission.
2.  provision for complete documentation of sources.
3.  provision for accessing any research E-Mail from the individual page in 
     GenWeb. 
4.  ease of finding an ancestral link in GenWeb.
5.  ease of finding and merging multiple occurrances of an individual
6.  ease of making corrections in pages in GenWeb

Please consider giving everyone direct access to GenWeb pages to facilitate 
rapid growth; i.e., no "middlemen" approving every move.  Give people the 
option to restore backup page versions, also. 

Please consider the use of separate marriage pages as demonstrated in the 
"Alternate Demo" at the above URL.  We feel that this approach will avoid 
duplicated lists of Children and be better suited to HTML-to-GEDCOM 
conversions later on.

Please consider a means of archiving and indexing all research E-Mail 
concerning an individual at a URL link on the individual page.  This way the 
work of today's researchers is not lost to tomorrows researchers. Provide no 
way to delete research E-Mail; this should keep people more polite in their 
exchanges.

We'd like your feedback to these ideas.  Feel free to blast these ideas if you 
have better alternatives.  Just present a good argument. 

We plan to write a GEDCOM to HTML converter and place a the 6000 name Richard 
Austin data base in GenWeb within two months.  This will give us a means to 
experiment with search/index techniques, and this data base will also serve as 
a great example of how to document sources!  By the way, we will look at 
frode's converter program to see if we can use his efforts.  If any of you are 
doing such a converter, please let me know this week.   

Looking forward to working with you all,   Bill Minnick, Cupertino, CA  


From UCSD.EDU!list-relay@netcomsv.netcom.com Sat Sep 10 21:14:18 1994
Date: Sat, 10 Sep 1994 23:58:05 -0400
To: svpafug@rahul.net (Bill Minnick)
Cc: genweb@UCSD.EDU
Subject: GenWeb2000:

Bill Minnick writes:

>We plan to write a GEDCOM to HTML converter and place a the 6000 name Richard 
>Austin data base in GenWeb within two months.  This will give us a means to 
>experiment with search/index techniques, and this data base will also serve as 
>a great example of how to document sources!  By the way, we will look at 
>frode's converter program to see if we can use his efforts.  If any of you are 
>doing such a converter, please let me know this week.   

I have already written such a converter, which is (at least) capable
of processing the 30K line "royal92" file from ROOTS-L.  I am sure that
it would be much less effort to process the "Richard Austin data base"
with my program than to rewrite another converter from scratch.
I am willing to help out with any modifications necessary to get the
database processed.

The C and YACC source for the converter can be obtained by visiting

	ftp:://cs.sunysb.edu/pub/TechReports/stark/HOME.html

I am accessible at "gene@starkhome.cs.sunysb.edu".

							- Gene Stark

From UCSD.EDU!list-relay@netcomsv.netcom.com Sun Sep 11 02:34:38 1994
Date: Sun, 11 Sep 94 11:09:38 +0200
To: genweb@UCSD.EDU
Subject: GenWEB - Global naming


What we need is a mechanism to look up persons in some location-independent way.
Perhaps some GenWEB name space server, so we could register our own personal
tags and the locations of our Web servers centrally.
This way, references to other people's trees would not have to rely
on the exact location of their base. If I move to australia and take
my base with me, I would only have to register with the GenWEB server
that my base would be temporarily unavailable, and then register the
new server upon arrival. I guess I would have registered a unique
GenWEB tag for my base based on the database's contents. Remember
that the base should keep it's tag even if it changes owner. And remember
that you may want to keep several separate bases. There shouldn't be
any problem mapping several tags to the same base, if you want to
use one base for now, but can foresee that you may want to split it,
and where you would want to split it.

I guess GenWEB URL's should be split in two parts: One global
tag, and one local implementation dependent part. Lookups could be
performed as
http://genweb.location/cgi-bin/tag-lookup/<global tag><separator><local part>

The global tag for my base could be any unique tag. Either by my own selection,
or just some automatically generated serial number. The separator character
should not be allowed in the global tag.

The local part can be anything that the provider wants, but it shouldn't rely
on things like the internal index numbers of the database. What happens if
you rebuild the base, or merge it with another one?

For those using HTML files, the local part should just be the last part
of the directory structure. If you have HTML files in a structure like
http://www.myserv/users/jones/genealogy/jones/albert.html
I would recommend registering //www.myserv/users/jones/genealogy with
the GenWEB server, and publish
http://genweb.location/cgi-bin/tag-lookup/jonestag%jones/albert.html
as the url for this person.
At the GenWEB server, tag-lookup should be a CGI gateway substituting
'jonestag%' with 'http://www.myserv/users/jones/genealogy' and issuing
a HTML line that makes the client load this page instead.

For people like me, using a gateway to do database lookups on the fly, the
published URL could be something like
http://genweb.location/cgi-bin/tag-lookup/jonestag%E32?Myindex
And I could have registered info making the server expand this to
http://www.myserv/cgi-bin/db-lookup/E32?Myindex
The index number used in this example should be a index number registered
within the data, not the private index numbers of the database application,
as you may have to rebuild the base, or move to another application, etc...


Pro's:
- Can be put together easily today.
- Shouldn't exactly overload the server

Con's:
- How do we handle the problem when/if the GenWEB server moves?
- If I want to change my own lookup methods (the private part of the tag),
  I should provide some compatibility gateway.


Birger

From UCSD.EDU!list-relay@netcomsv.netcom.com Sun Sep 11 20:33:23 1994
Date: Mon, 12 Sep 94 05:17:26 +0200
To: genweb@UCSD.EDU
Subject: Re:  GenWEB - Global naming

Birger A. Wathne writes:
>What we need is a mechanism to look up persons in some location-independent way.
>Perhaps some GenWEB name space server, so we could register our own personal
>tags and the locations of our Web servers centrally.

This is a problem which isn't limited to GenWEB, as just about any
document on the Web could exist on a number of unknown servers, and
one would like to access a particular document regardless of which
server it will be fetched from.  I think the easiest approach here
is to study the work already being done on URNs ("Uniform Resource
Names") by Tim Berners-Lee and others behind the WWW design.
Unfortunately I don't have the most current version of the URN
specification in front of me, but that could be worked out if there
is interest.  In any case, inventing a lookup scheme specifically
for GenWEB and no other documents would be a slight waste of time.
--
Anders Andersson, Dept. of Computer Systems, Uppsala University
Paper Mail: Box 325, S-751 05 UPPSALA, Sweden
Phone: +46 18 183170   EMail: andersa@DoCS.UU.SE

From UCSD.EDU!list-relay@netcomsv.netcom.com Mon Sep 12 08:14:22 1994
To: genweb@UCSD.EDU
Subject: Re:  GenWeb2000 Implementation
Date: Mon, 12 Sep 1994 07:31:04

RESPONSE TO:  Anders Andersson, Dept. of Computer Systems, Uppsala University
                                 Message of 11 September 1994

Anders:
Based on your observation, I'll design HTML pages for 640x480 monitors; higher 
res displays will have a lot of blank space to the right side.  This appears 
to be a shortcoming of HTML which we have to live with.  Do you concur?  

Apart form the details you commented on, I was rather looking for feedback and 
new ideas regardingthe more fundamental ideas I've proposed; i.e. the concept 
of keeping research mail with the individual information, and constructing 
GenWeb to be easily used by the masses - - without "middlemen" like ourselves 
controlling every addition ar change to the GenWeb files.  Please, your 
thoughts on these deeper matters.

Regards,  Bill Minnick

YOU WROTE:
  When a page in Mosaic begins

>........SILICON VALLEY PAF USERS GROUP
>HOME PAGE........

>it looks plain silly, in my opinion.  An <H1> element is appropriate
>here, but the periods supposedly added for the sake of symmetry (?)
>completely ruins any symmetry instead. ...
..............
>Maybe we need an H.T. Emily Webster set of guidelines regarding how
>(not) to write good (or bad) HTML code. ;-)

BILL MINNICK P.S:  Good Idea; please produce a strawman draft for review and 
criticism.