WWC snapshot of http://www.msi.umn.edu/paradise/history.html taken on Sun May 28 22:55:24 1995
Release notes are here
2-28-94

The History of Paradise Netrek

This document is incomplete. We need some input from Heath Kehoe, Mike McGrath and Bill Dyess as to what they think some major milestones are.

by Brandon:

November 92
Larry Denys & Brandon Gillespie (me) begin talking about the lack of variety with Netrek & how it could be much cooler.
January 93:
We've been talking for a while, Kurt has also had bunches of ideas. I begin downloading server sources and sifting Through various different ones. I run the flint source on paradise.declab.usu.edu.
February 93:
I finally manage to lure Larry and Kurt into the project, playing around starts. We toss the flint source and begin with the bronco source for both. Kurt does a majority of the client interior work, Larry is on the server. We discard the bronco client for the hadley client.
~March-April:
Paradise I has been extremely popular, primarily due to it being so different yet not complete chaos (some would beg to differ =)

The dates get vague here. Paradise I was the initial hacks after the initial hacks we withdrew the actual code into development and left Paradise I for itself.

Background on the situation at USU: Games are illegal in the labs we are developing in. However we had permission to run the server on one of them (paradise.declab.usu.edu) after hours. Dr. Larre Egbert, Computer Science professor was the original professor who ran the campaign against Netrek (as the no games rule was targeted towards Netrek). He claimed that stats had been taken and Netrek was taking up 60% of the available campus Network. (later legitimate stats have proven that if every student attending USU simultaneously played netrek it wouldn't even take up 5% of the available bandwidth).

Dr Egbert heard of the server we were running on paradise.declab.usu.edu and began pulling strings again. I received a letter quickly written explaining the situation (which was very tense), within 15 minutes the currently running game was nuked and all code, files, etc were tarred backed onto floppy and removed off the system, and we went on spring break.

After spring break I began looking for alternate sites for the server, as we were about to release paradise II. Mike Lutz and Heath Kehoe ran it for a brief while alternate to netrek.icaen.uiowa.edu. Rob Forsman stepped in and started up the first full time functional paradise since the initial paradise.declab.usu.edu had been shutdown at mole.cis.ufl.edu, later renamed to netrek.cis.ufl.edu.

We found another machine (recumbent.declab.usu.edu) through the graces of Bob Bayne, Manager of Academic services at USU. Paradise II was released.

June ~ current:
Kurt dropped out of the project, a development team was essentially formed between others (formed around april actually) With Rob and Larry at the core. Larry left the team when summer started.
December 93:
recumbent.declab.usu.edu can no longer handle the usage of the server (it was a weak machine). Charles Swenson, (aka alfven) a professor in the Engineering Department at USU offered a home for the server, with more memory.

pippin became the new home.

so basically:

by Bob Forsman

Arctica was named after an Empire country I played in a once-per-day Australian game. It was first brought up on mole.cis.ufl.edu. The sysadmin was gracious enough to give me a cname: netrek.cis.ufl.edu which came in handy when he relocated it to floater.cis.ufl.edu, which was a consultant-only sparc. Later he decided to declare floater a lot more unstable and changed the cname to beach.cis.ufl.edu where it lives today. Beach was upgraded to 2 processors late February 1994.

Paradise 1

March 19: Bob Forsman retrieves the 1.05 source code.

March 23: Bob asks to be added to the paradise mailing list because he now hosts paradise's old database.

Paradise 2 (1.99x)

These dates are from files I have in my directory. They probably lag the source release by a little bit since I didn't maintain the source archive back then.

Paradise 2.01

The 2.01 server was released probably in late June, or early July. It was called 2.01 because the 1.99x series was accidentally released under the name 2.0 at one point. The 2.01 server was released some time before July 15th. I can't set a better date because the earliest patch is dated July 15th and the full source tar is dated August 19th (implying that it was compiled/altered well after the release). The client was released before July 2nd. Similar dating problems apply to it since the full source tar was the only base archive of 2.01 client source.

First prototype for the tournament queue was developed at Arctica and announced on 4th of July. The client side was released as 2.01 patch 4 on 9th of July but didn't completely work until 2.01 patch 6 released on 17th of July. Server 2.01 patch2 and 3 were released the same day and were the first release of the server-side TQ code.

The league server was first tested in a game at Arctica on 24th of July. Brandon in his .Lynx incarnation was the primary planet scummer for the Orions while Muckraker (an Admiral who kicked my ass all over the place when I played sturgeon-style) made an appearance to help out Hammor's team. The first league server was released as 2.01 patch 5 on 27th of July.

Paradise 2.1

2.1 was released around August 10th. Shortly afterwards Brandon suggested giving our clients an expire time.

The high scores program, which prints out who's been playing the best recently, was released in August.

The idea for game parameters was introduced in late July. The first fragments of code were released on the 15th of August in server 2.1 patch 2.

Brandon created the info server on 17th of August. It ran on recumbent on port 3000. It was a normal netrek server that was closed all the time and had information in the MOTD.

Mike McGrath's fighters are incorporated into the mainline source by Robert Forsman on August 29th in server 2.1 patch 6. Fighters were determined to be extremely unbalanced and have been shelved. The code is still in there, though.

September 13th saw a posting from Adam Harrison saying (among other things) that he and Shawn Bairds had incorporated sturgeon upgrades into his copy of the Paradise source. Brandon had a coronary. This mod was quickly flamed out of existence, with no hard feelings towards the misguided implementors.

September 24th: Bill Dyess sends his first client patch to Robert Forsman. Setting a pattern for future patches from Auburn, it was enormous. It incorporated a lot of ideas from BRM, but (in typical Paradise style) better. Its final form was released on October 7th as 2.1 patch 2.
The interval between 2.1p1 and 2.1p2 was ~37 days, the longest interval between client patches to date. Patch 2 was followed by patch 3ufl by a couple of hours.

Oct 1: Another record hiatus of 31 days between server patches is ended with the release of server patches 8ia and 9ufl. September was just a dead month.

Paradise 2.2

The 2.2 server baseline was released on October 29th by Brandon. The 2.2 client baseline was released on November 5th by Bill with another huge wad of way-cool hacks including macros.

Late November '93 the Paradise League Council (PLC) actually starts doing things. Members vote on which features will be part of standard Paradise and which will be off by default. November 30th, Brandon posts a schedule of league games. The teams are: .D.O.T.S., Future Business Majors, <this space for rent>, and The Hammor Party.

December 11th saw the first league game: <this space for rent> -vs- The Hammor Party. The controversial Light Cruisers had been enabled by the PLC and Mike Lutz hadn't heard about it. Mike was so pissed he dissolved <this space for rent> and refused to play Paradise League "in this jello for rules invirernment". Heath Kehoe (a teammate of Mike's from the same university) less than an hour afterwards declares that he is tired of his ideas being ignored by the other developers and resigns from the PLC. The other members are somewhat confused by all this.

December 12th sees the first prototype observer mode at UFL's experimental server. It is relased in server 2.2 patch 8 with many other enhancements to the league server.

December 16th sees the integration of Eden's PT boats and Utility vessels into the mainline source.

December 28th sees the release of the client-side observer code.

Hasta, 1993. Happy New Year 1994
January 6th: client 2.2 patch 8 adds the kitchen sink in response to Andy McFadden's (?) sarcastic comment on rgn.

PLC has its first electronic meeting on January 10th at a MOO running on Recumbent. These become a weekly affair. Those who attend decide the fate of paradise. Those who don't are not allowed to complain.

January 17th: PLC ratifies server 2.2 patch 11 as the official league source for the Winter 1994 season. They also ratify a rulebook and shortly afterwards post a new schedule to begin the Winter '94 season. Teams are: Future Business Majors, Make Money Fast, and The Hammor Party.

January 19th: Mike McGrath integrates Broncoish capability into the server baseline as 2.2 patch 12.

Paradise 2.3

March 3rd: Brandon prepares a 2.3 server baseline and hands it off to Rob Forsman. A few bugs are discovered in the implementation of a new non-exponential army growth scheme and are fixed before the source is indented and tarred. It is announced that afternoon.

March 10th: 35 days after patch 10ufl, client hackers fire off a quick round of patches. This sets a new record for delay between client patches. Client baseline 2.3 is then released. It incorporates more COW features to help make paradise more popular with the normaltrek crowd.

March 11th: Joe Rumsey is finally added to paradise-workers. He had been hacking an Amiga client for paradise outside the umbrella of paradise-workers but wanted his changes to become part of the mainline source.

March 23rd: Joe Rumsey provides the first amiga binary for paradise to be archived on ftp.cis.ufl.edu.

April 4th: Larry Denys is sick of twinks on pippin. He can't seem to get the preexisting clue checking code to work the way he wants so he writes a quick hack of his own. The PLC responds by upgrading the preexisting clue checker. The concept of clue checking is the subject of quite a bit of discussion on the mailing lists.

April 7th: UFL ends a record streak of server patches starting with 2.2p17 (Feb 28) and ending with 2.3p5 (Apr 3) by releasing 2.3p6usu,ufl, which incorporates some designs from Brandon&Co. Other hackers have not had to release patches because they haven't released anything broken.

The saga continues: Part 3

May 6th - August 2: Server patches 7 through 13 are released from various sources, which will be documented here later. Client patches 2 through 9 are released as well; development on a new top secret client begins.

August 3 - Bill releases server patch 13 incorporating Warp Zones, improved floating point performance, and some hacks to make it compile under Linux.

August 3 - August 15: Rob and Brandon announce that they are leaving paradise development. WWW and ftp site are transferred to Bob Glamm. Current UFL WWW pages & ftp site to remain active at least through the opening of Bob G's WWW pages & ftp site.

August 20 - After a painful experience with diff's whitespace option, Mike releases server patch 14; this patch includes a variety of fixes from a variety of authors, including the insidious planet off-by-one bug, wormholes in galaxies 1, 2, 3, and 5, wormhole effects, and new SYSDEF options to control wormholes & the galaxy metal planet layout. Contrbutors: Mike, Bill, and Rob.

August 23 - Bob G releases server patch 15; this patch is essentially a port to SGI machines.

[RG]

[Paradise]

Paradise Netrek Hackers / <paradise-workers@reed.edu>