Copy Link
Add to Bookmark
Report
Doom Editing Digest Vol. 01 Nr. 163
From: owner-doom-editing-digest
To: doom-editing-digest@nvg.unit.no
Subject: doom-editing-digest V1 #163
Reply-To: doom-editing
Errors-To: owner-doom-editing-digest
Precedence: bulk
doom-editing-digest Friday, 17 February 1995 Volume 01 : Number 163
Re: almost objective editor comparison
Re: almost objective editor comparison
Re: almost objective editor comparison
Re: node generator comparison
Re: Nodebuilding times
Re: DEU 5.3 Beta
Re: DEU 5.3 Beta
Re: DEU 5.3/editor comparison
Re: almost objective edit
----------------------------------------------------------------------
From: fenske@rocke.electro.swri.edu (Robert Fenske Jr)
Date: Thu, 16 Feb 95 12:23:36 CST
Subject: Re: almost objective editor comparison
>> 2) Node generator is 100% reliable now. I challange anyone to find a map
>> it cannot build nodes for. It is also 'the fastest node generator on
>> the face of the planet!' Since WARM made this claim of it's node
>> generator, and mine beat it, I guess it's my title now.. :)
>>
>> Node generator build times for E1M1
>>
>> WARM DMapEdit
>> 11 seconds 9 seconds
>>
>> Anyone else want to compare theirs?
>
>DeePBSP (although I almost hate to admit it) compiles E1M1, including the
>blockmap, in 2.97 seconds.
Ahh, I was wondering how long it would take before someone else
would lay claim to "fastest node generator on the planet". I see that I
should have left the "quiet" option in, which turned off the screen I/O
during the nodes and reject building. WARM is significantly limited by
the screen I/O speed and runs about 1/3 faster without the screen output.
Anyway, if DmapEdit's builder is faster on a bunch of WADs then I suppose
you get the title. From your times I'd say you have a 486DX/33. I've never
heard of DeePBSP, but that 3s time must be from a Pentium. WARM certainly
flies on my P5-90.
Robert Fenske, Jr. rfenske@swri.edu Sw | The Taming the C*sm*s series:
Electromagnetics Division /R---\ |
Southwest Research Institute | I | | "The Martian canals were the
San Antonio, Texas USA \----/ | Martians' last ditch effort."
------------------------------
From: fenske@rocke.electro.swri.edu (Robert Fenske Jr)
Date: Thu, 16 Feb 95 12:45:19 CST
Subject: Re: almost objective editor comparison
>One thing I'd like to point out, though, is the reliablility and
>optimization issue. I could probably write up a node generator that
>builds node in under a second, and would probably work fine, but I think
>that they should strive for best cases of least splits/even balance.
>Doing this is what takes the most time. If DeePBSP is running that fast,
>I for one wouldn't trust it to be optimizing the bsp tree very well. My
I very much doubt ever getting a running time < 1s (using a 486/33)
unless you wrote chunks of code in assembly. WARM spends about 70% of
its execution time deciding on which side of the partition line a seg belongs;
because this decision has to be made so many times. Even a 3x speed-up in
the decision routine wouldn't shave that many seconds off the running time.
Now as you suggest having a very crude partition line selection process
would greatly speed things up, but the node tree wouldn't be optimized at
all.
>code uses math is about it's fastest form, thanx to help from Robert
>Forsman (math god :) and the guy who wrote WARM even. I learned about the
>Ax + By + C = 0 form of a line he was using in his earlier project RBM. I
>have a hard time figuring out his code very much, but I think that we both
>use pretty much the same ideas and methods actually. Probably the same
>math formulas anyway.
Yes, geometric and algebraic manipulation can go along way in
eliminating unnecessary computation. I think CS degrees should require much
more math than they do now, but that's another story (I have CS and physics
degrees: in physics you get your face full of math).
Robert Fenske, Jr. rfenske@swri.edu Sw | The Taming the C*sm*s series:
Electromagnetics Division /R---\ |
Southwest Research Institute | I | | "The Martian canals were the
San Antonio, Texas USA \----/ | Martians' last ditch effort."
------------------------------
From: setc@together.net (Ross Carlson)
Date: Thu, 16 Feb 1995 14:01:27 -0500
Subject: Re: almost objective editor comparison
>> > True. It also used to be hard to figure out, too, I think maybe. Dck and
>> > Edmap came along later, and people talk about them now pretty often.
>>
>> Well, I just tried DCK2.0, and I'm _not_ impressed: even I could make a
>> better copy of DEU than that.
>
>DCK is by no stretch of the imagination a "copy" of deu. Of course it has
>its similarities - every doom editor can be compared to another.
>
>I guess this attitude is the same one that prompted you to, unprovoked,
>hack DCK in r.g.c.d.e when you really didn't know what you were talking
>about. (for reference, you said "DCK users wouldn't know how to fiddle
>with sector refs if they needed to" - probably one of the most obvious
>operations in the editor. But I doubt you knew that at the time, because
>you "just tried DCK 2.0" as of now.
>
>It's a shame statements such as "even I could make a better.." are so easy
>to hide behind, or I'd call you out on it. Doubtless you'd excuse
>yourself as a result of your busy schedule or sudden indifference to the
>subject.
>
>I apologize to the readers of this list for replying in kind to this
>message.
>
>- Ben
Right on Ben!
Seems to me, that DCK 2.0 is seen as one of the better doom editors out
there. Don't know what this guy was thinkin' - probably wasn't!
- -Ross
------------------------------
From: fenske@rocke.electro.swri.edu (Robert Fenske Jr)
Date: Thu, 16 Feb 95 13:27:39 CST
Subject: Re: node generator comparison
"Jason Hoffoss" <hoffo002@gold.tc.umn.edu> wrote:
>Hmm, maybe I could write up a little program to test the optimization and
>reliability of bsp trees, so we can rate node generators better. Ya,
>think I will. :)
What are you thinking here? Some criteria I can think of are
fewest splits, fewest nodes, depth of node tree, width of node tree,
# left branches, # right branches, and various ratios of these measurements.
One thing I have been interested in is computing the optimal node
tree for a given level and how close (measured in some fashion, perhaps
subjectively) a node builder can reasonably get to this optimal tree. But
I have considered that finding the optimal tree would be computationally
prohibitive. I say this mainly because I don't think that simply choosing
the best partition line for a given segs set will produce the optimal tree,
e.g. a particular "best" partition line may force a quite unbalanced tree
farther down. So I think this means that a whole pile of trees would have
to be computed and then compared.
Robert Fenske, Jr. rfenske@swri.edu Sw | The Taming the C*sm*s series:
Electromagnetics Division /R---\ |
Southwest Research Institute | I | | "The Martian canals were the
San Antonio, Texas USA \----/ | Martians' last ditch effort."
------------------------------
From: bmorris@islandnet.com (Ben Morris)
Date: Thu, 16 Feb 95 13:01 PST
Subject: Re: Nodebuilding times
S>Ben Morris wrote:
>>
>> DeePBSP (although I almost hate to admit it) compiles E1M1, including the
>> blockmap, in 2.97 seconds.
>>
>> Crazy, eh?
>>
>It certainly is. And what did WARM and DMapEdit get, on your P5-100? :)
That's on a 486 dx2/66 :)
BSP 1.2x scored 7.5 seconds.
------------------------------
From: bmorris@islandnet.com (Ben Morris)
Date: Thu, 16 Feb 95 13:01 PST
Subject: Re: DEU 5.3 Beta
djsd100@hermes.cam.uc.uk (David Damerell) wrote:
D>DCK seems to attempt to insulate me from the existance of sidedefs - I
>never saw a sidedef number listed - can I attach several lines to one
>sidedef?
No, you can't. You don't even NEED to - there's no advantage to this
apart from the fact that it might make edits a little faster. In the
end, though, it makes no difference to the functionality of the level.
D>Also DCK seems to have a very coarse display - DCK in 640x480 looks like
>DEU in 320x200 to me.
I assure you this is some kind of subversive plot. 640x480 is 640x480.
D>I downloaded DME 4.0 (beta 2): the docs would seem to imply that the
>floor-ceiling heights are represented more coarsely than in DooM, and
>that DME will sometimes flip linedefs because it thinks I have something
>wrong. Maybe I _want_ it wrong. This is, IMHO, not complete control.
Maybe you want it wrong. Can you give me an example?
D>Ah. You said 'all far more sophisticated than the 5.3 beta is - FAR
> ^^^
>more'. Hence I presumed that 'all' included all editors. If you define
>'all' to exclude Windoze editors, that's a funny 'all'. OTOH, you don't
>seem to have said what other editors have that DEU doesn't.
(The following is all correct to the best of my knowledge, but if I've
missed a feature for either EdMap or DME - sorry.)
DCK and EdMap:
- Series (> 1) of Styles for decorating sectors/lines
- Automated set-tags function
- Search / Replace textures
- "Run DOOM from editor"
DCK and DMapEdit:
- Graphical thing display.
- Auto-merging of lines. Not just congruent lines, but lines that
overlap. DCK includes diagonal merging.
- Multi-texture viewer
- Automated texture placement
DCK, Edmap AND DME:
- Automated stairs
- Thing display masks
- Complete grid control (not just "G" "G" "G" "G" "G")
That's not including features that one of the editors has over all the
others.
- - Ben
------------------------------
From: l-sieben@MEMPHIS.EDU (ulasieben)
Date: Thu, 16 Feb 1995 15:37:29 -0500 (CDT)
Subject: Re: DEU 5.3 Beta
>On Tue, 14 Feb 1995, Ben Morris wrote:
>> > > > Finally! Its out! YES! Beta 1 is great...it just has those usual
>> > Could someone be able to tell me as to where I could ftp this file
>> > from, as I thiknk a renewal of wad editing is due.. :)
>> Out of interest, have the DEU team been keeping up-to-date with any of the
>> latest versions of the other editors? They're all far more sophisticated
>> than the 5.3 beta is - FAR more.
>
>That, of course, is a matter of opinion. I'd say that DEU 5.21 is
>indisputably superior to most other editors, although DCK looks like it
>may be quite cute. So unless 5.3 is a big step backwards...
>
>But that's just my opinion. OTOH, yours seems a lot more extreme - you
>think the Windoze editors are better than DEU?
>
>David Damerell, GCV Sauricon. djsd100@hermes.cam.ac.uk Trinity, Cambridge
I agree. Most editors coming out now are for the newbies just starting
out with
D2, and have no clue about WAD structure. Hence, DoomEd fits into a niche
quite nicely. DCK is not that bad, it just has to mature a little more (in
texture
WAD area), but looks good overall. DEU & Edmap are popular editors because
they have
been there a while, and have a following. Waded is in the same category as
DoomEd,
because it is easy to use, (except no docs). Not sure about DoomCad, but its
supposed
to be pretty good. DMapEdit is a powerful editor, and is gaining. What am I
getting at?
All the editors out there have their own niche, kind of like in ecology.
Vary rarely in
nature will you find 2 species competing in the same environment. One or
both usually
adapt. The same thing is occuring now in Doom editors. At first, there was
DEU and
DMapEdit, I think. These were mainly for the people that knew all about Doom
a year
before it came out. Then, stuff like EdMap started appearing to bridge the
gap. Now,
the programmers of new editors are aiming for the new market, the Doom 2 and
Heretic
people. These people don't know and probably won't ever learn the WAD structure,
becuase they have no need to. That is what makes this all so great:
diversity. DEU,
EdMap, and DMapEdit are for the people that want to make serious wad with
all kinds of
little tricks in them, and want raw power. The stuff like DCK and DoomCad is
for those
that are intermediate in knowledge and need for power. DoomEd and Waded are
made almost
exclusively for the new WAD maker, the kind that just makes a few DeathMatch
levels for
him/her and a few friends, and never get released.
As for the debate over the DEU 5.3 B1 release, I think it has potential.
First, the
new dialog boxes will help out immensly in the speed department. Supposedly,
DEU 5.3
has a randomizer in it, a kind of gee wiz feature. Of course, the D2 and Heretic
support are major works. However, much of the work is unseen to the user
yet: the readme
says DEU will soon be able to edit multiple maps at the same time and
cut/paste between
them. Not many editors can claim that yet (DMapEdit comes to mind, though).
Also,
the core has been rewritten to speed porting to other platforms. That take a
*while*
to do, and will help out in the end because of greater availablity (sp?).
Keep in mind this is a BETA version. Betas typically are underdeveloped.
Just my $.02 on the matter.
-- Evil Genius (Jimmy Sieben)
------------------------------
From: Piotr Kapiszewski (Kapi) <kapis-p@grinch.cs.buffalo.edu>
Date: Thu, 16 Feb 1995 16:49:58 -0500 (EST)
Subject: Re: DEU 5.3/editor comparison
Raphael.Quinet@eed.ericsson.se said:
> Hello everybody...
> checking my ftp server to see if a new version had been released. On the
> other hand, lots of people were complaining because they saw nothing new for
> DEU. So I released a beta version for them, but without announcing it.
> 8 minutes after I put the file on my server, someone had downloaded it. 20
> minutes later, it was on ftp.cdrom.com. People have no lifes. :-) This is
Hahahaha I love it. It reminds of a similar thing when I was working on
Web stuff for DoomGate from home. Yahoo had a probe run though my system and
my box on a puny 14.4 ppp link became one of the top sites for access. You
can imageine what happens to the link like that if two people try to hit it
with http viewer. People really have no lifes... but then again who are we
to talk about it....
Sorry about this being a bit off topic but I though it was really funny.
Keep up the great work Raphael.
- -Kapi
- --
PGP key available from doomgate.cs.buffalo.edu | Piotr Kapiszewski
URL at http://doomgate.cs.buffalo.edu/~kapis-p/ | kapis-p@cs.buffalo.edu
------------------------------
From: matt.tagliaferri@pcohio.com (Matt Tagliaferri)
Date: Thu, 16 Feb 1995 18:57:00 -0500
Subject: Re: almost objective edit
DO>> Well, I just tried DCK2.0, and I'm _not_ impressed: even I could make a
DO>> better copy of DEU than that.
DO>DCK is by no stretch of the imagination a "copy" of deu. Of course it has
DO>its similarities - every doom editor can be compared to another.
DO>I guess this attitude is the same one that prompted you to, unprovoked,
DO>hack DCK in r.g.c.d.e when you really didn't know what you were talking
DO>about.
I had a feeling this would turn into a "my editor's better than yours"
argument before long....
matt tag
(humble) author: DoomCAD
- ---
þ OLX 2.1 TD þ Intel Inside: Smoke on the outside
- ---------------------------------------------------------------
PC-Ohio PCBoard pcohio.com
The Best BBS in America Cleveland, OH
DATA: 216-381-3320 FAX: 216-291-2685
- ---------------------------------------------------------------
------------------------------
End of doom-editing-digest V1 #163
**********************************
