History of the RailServer

The RailServer has a long and sometimes turbulent history. It is one of the largest and most important projects I have worked on until now, and there's a lot of "personal touch" in it. I learned a lot. And here's the story...
December 1993
My account is created on one of our computing centre's Unix workstations, and I begin to explore the Internet.

early 1994
I start reading USENET newsgroups, among them soc.culture.german where I try to help foreigners who ask questions about rail connections in Germany.

Saturday, May 21st, 1994
I create my own WWW home page with a section on railroad schedule information, encouraging people to eMail me their specific questions. I ask Robert Bowdidge, who already has a large web page about railroads, to put in a link.

Wednesday, June 8th, 1994
I have completed a program which automatically reads rail queries out of my mailbox if they have a specific format and answers them using a schedule software available on disk. I send an annoucement to some of my friends and ask them to test it.

Thursday, June 15th, 1994
Andreas Ley at the computing centre creates an eMail alias for me, rail@rz.uni-karlsruhe.de, so that I don't have to sort the queries out of my personal mailbox. He has also written a small mailto script that allows me to construct a WWW interface where people can enter their queries (the response is, of course, sent back by mail).

Monday, June 20th, 1994
The automatic service works fine. I have my computer dial into my university account twice a day to fetch and process the mail queries. An announcement is sent into all newsgroups that deal with travel in Europe or Germany.

Monday, August 15th, 1994
The RailServer is mentioned in the German news magazine "Focus".

Thursday, August 25th, 1994
The RailServer has answered 9239 queries from Germany and 3218 queries from foreign users.

Saturday, August 27th, 1994
I write a letter to HaCon, the company selling the schedule software the RailServer uses. I suggest a meeting and a discussion about a professional version of the RailServer with their support.

I receive a friendly answer, and in October a meeting between HaCon, someone from the "Deutsche Bahn AG" (the German rail operator) and some compunting centre people takes place - without tangible results. Meetings!

June 1995
One year old, the RailServer now serves up to 1000 queries a day. My 486 PC at home needs up to two hours per job to compute the answers.

July 1995
A Frankfurt/Main based company offers online schedule information, but many users don't like the fact that you have to register with them at "germany.net" first. DBAG opens their own web server, but they don't have schedule information (a link to the RailServer gets inserted later).

September 1995
The computing center provides me with a PC that wil be installed there and serve queries around the clock. What a pity, I got used to impressing people at home by telling them: "This computer processes queries from around the world..."

Sunday, January 22nd, 1996
HaCon asks if I was willing and able to create an online schedule information for them. (One day earlier, and this would have been a fantastic birthday present.) Of course I am! Intense cooperation begins.

June 1996
I celebrate the 2nd birthday of my "baby". The RailServer now answers up to 2,000 queries a day and has served half a million altogether. At the same time, I know that it's best times are over, because the online service I created for HaCon is now publicly accessible. Nevertheless I keep the RailServer running because no one knows if DBAG will allow HaCon to run their server forever.

I assume the post of "WWW admin" for the HaCon service, answering all the user questions (and flames).

Tuesday, May 6th, 1997
The RailServer has shrunk to about 500 queries a day. For a moment, it seems that it may become important again as DBAG tries to close down the HaCon service (see special report). Fortunately, that doesn't happen; the RailServer with its somewhat outdated technology would not be able to cope with about 20,000 queries a day that HaCon serves.

September, 1997
I decide not to renew the RailServer schedule data when it runs out at the end of the month. Thus ends a project that has lasted for more than three years and never cost or earned any money. The RailServer has nevertheless been of use to many people, including myself because there were commerical spin-offs. I am especially happy about the many foreigners who used the RailServer to plan trips they might otherwise have made by car.

An automatic answering script takes care of the mail address rail@rz.uni-karlsruhe.de, and I am trying to eliminate the RailServer from all the search engines and hotlists. I get a lot of eMail thanking me for the service.

Melancholy creeps up inside me, but I keep saying to myself that the RailServer has done its duty and that the small number of queries in the past weeks really doesn't justify the work of keeping it alive.

End of March, 1998
The HaCon online schedule service serves over 100,000 queries on some days. I quit my job as "WWW admin" to concentrate on my thesis: a Java rail schedule interface...

End of May, 1998
My thesis is finished, my studies are done - I hope the rail people will survive on their own :-)

T H E   E N D

  Frederik Ramm, 2002-07-12