Millenium database

Re: Talk.Bizarre Second Millennium database help needed
Message-ID: <7baqve$t30$1@ruby.ansuz.sooke.bc.ca>
   
In article <36d8c256.1959387@news>, Eric O'Dell <eodell@pobox.com> wrote:
>This will be assembled into a browseable, searchable web database with
>lots of nifty features. As soon as the lights go back on after Y2K,
>I'll burn it onto CDs for anyone who wants it, free of charge.

Beware!  I did this for alt.kids-talk.  Let me tell you what you're in for.

First of all, there are copyright issues.  Not everyone will be happy with
allowing you to copy their work, especially if you permit redistribution
of the disc.  (Paraphrase from one akt poster: "I don't want my picture
and edited-down clips from my postings, to appear in a 'strange people of
Usenet' magazine article!")  On the other hand, some people will WANT
their material to be widely distributed, and object if you do anything to
limit distribution.  Every amateur lawyer will have an opinion on what
language you should put in the license, and the philosophers will question
your motives, wondering whether you're doing the project for personal gain
or the good of the group.  All these people will make a lot of noise and
ultimately be unable to stop you from doing whatever you feel like doing.
But it can be quite unpleasant along the way.

Then there are the privacy issues.  Some people will want you to exclude
all their postings from the disc.  They, of course, will be the people who
changed addresses repeatedly during their time here and don't remember
exactly what all their addresses were.

So you get your list of postings together and find that there are holes in
it.  Now you have to negotiate with the people who have archives, all of
whom will put conditions on how you may use the material they have, if
they even allow you to have it at all.  (There are three major archives of
alt.kids-talk; one operator expressed misgivings about the project but
decided to trust my judgement; the second refused to give me one byte
because he thought the past should not be remembered; I started the third
archive myself, against future need.)

You spend two months writing the indexing software.

You accumulate all the postings on your hard drive, find it's too small.
Buy a bigger hard drive.  You start processing them with Perl scripts,
thrash your machine into the ground, buy a memory upgrade.  By the way,
don't even *think* about attempting this under Windows.

You spend a month chasing bugs in the thread-indexing system.  Meanwhile
you're soliciting contributions of biographic information from regular
posters.  ALL of them promise to send you something.  NONE of them send
you anything.  You extend your deadline.

A former regular rejoins the group, hears mention of your project, and
restarts the "hey, is this a good idea?" flamewar from the beginning.

You build a test CD image, and your premastering software goes into an
infinite loop.  Obtain the source code.  Debug it.  Discover that if you
want it to be readable on a Mac, the Mac filesystem format has a limit of
65536 allocation units per filesystem.  You want to store 90000 files on
this disc (note, for talk.bizarre the number would be somewhat higher,
this was just a couple years' worth of akt because the guy with the really
old archives refused to let me use them, see above).  Something has to
give.  You end up building the disc as ISO with extensions for Windows,
Mac, and Unix.

The day before the "really real final" deadline, three people send you
biographies.  That same day you decide it's time to do something about the
"bus error" messages you've been getting with increasing regularity.  Many
hours of troubleshooting later you figure out that it's the RAM upgrade
you bought.  Whose warranty expired a week before.

You build the final CD image.  Forget a file, have to rebuild it.
Discover that your operating system, under some circumstances, when it
mounts an ISO image in loopback mode cannot read the last file in the
image.  Add a dummy file to prevent this from breaking your MD5
self-checksum script.

Test the disc on every computer you can find.  It works perfectly on all
of them.  Pat yourself on the back.

Burn a stack of discs and send them off to people who requested them.
Burn a few extras for people who might want them later.  Explain to
several people that no, you really do have to have any payments in *your
own* country's currency.  Leave the job where you had access to the CD-ROM
burner; hope that you have enough extras.

Suddenly get a pile of orders; fortunately, you DO have enough extras.

Weeks later, hear from someone who tried to read the disc on a Mac.  It
doesn't work.  Much debugging later, determine that your image doesn't
work on any Mac, except the one you tested it on originally.

Years later, someone sues you.  (That didn't happen to me, but it will to you.)
--
Matthew Skala  Ansuz BBS  (250) 472-3169  http://www.islandnet.com/~mskala/

                            GOD HATES SPAM

Update

Subsequent to my writing the above, someone actually did sue me.

Comments

Copyright © 1999,2001 Matthew Skala
Updates to this entire site:
Updates to this category (teebee) only: