Hard Light Productions Forums
Modding, Mission Design, and Coding => FS2 Open Tools => Topic started by: Hades on May 02, 2010, 05:14:35 pm
-
I'm just curious but is the source code for modview available somewhere? Or does anyone have it on hand on their hard drive, etc?
If this is posted in the wrong place then please move it, I wasn't sure where to post it.
-
Yes, it's available somewhere.
-
Any idea where? I'm curious because someone asked me if I knew where it was in a discussion about the program. I guess he wants to try and update it to today's standards (glow-points, etc) because he and I both have negative thoughts about PCS 2.
-
Excellent idea.
And I found the code (http://www.descent-network.com/cgi-bin/download.cgi?file=/ddn/sources/modelview/modv32b5src.exe&size=1064&desc=MODELVIEW32_Source) by looking on the Descent Network site.
Also, read this thread (http://www.hard-light.net/forums/index.php?topic=11034.0). Search is A-1 SUPAR.
-
you would probably also have to support shield collision as well for full compatibility with current pof specs. pmf format support would be good too :D
-
Okay, to revive this piece of software a bit, I have put in some work to get it to compile on VS2010.
It compiles, but beware of any lingering bugs. The author made a few very, very wrong assumptions about VC's scoping behaviour, and how it wouldn't be changed from VC6's buggy implementation. I may have introduced some errors where none were before.
Get it here: http://blueplanet.fsmods.net/E/ModelViewSource.7z
EDIT: ModelView project on Google Code, started by yours truly: http://code.google.com/p/modelview/
-
Thank you very much.
-
OMG is this really happening!?
:nervous:
Don't wake me!
-
It's only happening if other people can come forward to start helping (and yes, bug reports and feature requests etc ARE helping, so go to the Google Code page I linked to and start entering them!).
-
I hate to be mister negative here but doesn't this project divide up already thin resources? Having the source is great but why try to revive a long dead util? Would it not be better to implement the best features into PCS2?
-
ModView has the advantages of stability and user-friendliness, both of which are very important. You could just as easily make the case that people should implement the best features of PCS2 into Modelview.
-
Either way having 2 different projects is a waste of programming assets. Maybe the real goal should be a new conversion program that is the best of both worlds. PCS2 already has the advantages of dds support, high poly support, glow points, etc.
-
A brand new program would be precisely the wrong thing to do; why start from scratch when you can build upon something that already exists?
The consensus seems to be that the PCS2 codebase is very hard to work with, and this is complicated by the facts that 1) the PCS2 code repository isn't available for new members; 2) the most recent PCS2 code isn't even in the repository. In contrast, the Modelview source code is freely available, well organized, and well documented.
In view of these facts, focusing on Modelview would seem to be the better choice.
-
I didn't mean start from scratch. Start with either PCS2 or modelview (probably PCS2 as it's more up to date with features and compatibility) and start a new project that pulls the best into one program.
-
With all due respect, FUBAR, but you might have a point if anyone was actually working on PCS2. There is no indication of that beyond a few posts around here saying "I'll get around to it".
And after sticking my head into the modelview code, I'd have to concur with Goober's assessment that it is easier to work with. For one, it can be made to compile on modern systems rather easily. Second, it does (at this point) not require any external framework.
-
And after sticking my head into the modelview code, I'd have to concur with Goober's assessment that it is easier to work with. For one, it can be made to compile on modern systems rather easily. Second, it does (at this point) not require any external framework.
I assume that means it only runs on windows?
-
Yes, since it is a MFC application. Changing that is on the To-Do list, but I'd rather have a non-memory-leaking, updated version running on Win before tackling that.
-
That's probably one of the stronger arguments for working on PCS2 instead, we've already got support for most of the formats we need, it's already a wx application, etc. Moving from MFC obviously takes some time, as we've seen with trying to get FRED away from it. I'd rather see efforts towards a cross platform app.
-
Sure, PCS2 being a wx app already has its benefits. Ideally, a merged application combining PCS2's functionality with ModelView's UI would be preferred.
But how long are we going to wait for development on PCS2 to pick up again?
-
I like PCS2's UI...
I just don't like its error messages.
-
The only thing keeping it from being worked on currently is the difficulty getting it to compile it looks like.
-
I'd just like to throw in my two cents to support ModelView.
I personally cannot stand working with PCS2, and I would love an alternative. I am not a coder myself, but if it truly is more user friendly, and works natively with current windows machines, then I would love the **** out of it. PCS2 is a bear to work with, and it's UI leaves a lot to be desired, as an artist rather than a programmer, I would love to see a more user friendly model importation utility.
-
I'd just like to throw in my two cents to support ModelView.
I personally cannot stand working with PCS2, and I would love an alternative. I am not a coder myself, but if it truly is more user friendly, and works natively with current windows machines, then I would love the **** out of it. PCS2 is a bear to work with, and it's UI leaves a lot to be desired, as an artist rather than a programmer, I would love to see a more user friendly model importation utility.
^ This.
-
I blatantly copied PCS2's source into my own SVN. Checkout URL is https://modelview.googlecode.com/svn/branches/PCS2, if you want to work on it, drop me a line and I'll set you up with commit access.
Note that, as present, it is uncompilable, as there are files missing from the PCS2 CVS. It is also lacking collada support. If someone has the source for that, I'd be interested in it.
-
Double post for great justice!
If you are interested in helping out, and have a working copy of modelview installed please download this exe, and replace the old one with it:
http://blueplanet.fsmods.net/E/MODVIEW32.exe
It's my current version of it, and it needs stress-testing. In terms of functionality, it's the same as the last public release, but I bumped some limits (A LOT). Please check for the presence of legacy bugs, and make note of any new issues you find.
Now, I know this doesn't have the functionality to build full FS2 models (Path editing, for example, is noted for its absence). It'll be coming.
-
No luck here. Splash screen and hour glass. XPsp3
-
It's a shame there's even limits and we can't just convert those things to vectors...
-
We probably can. And if "we" was more than "me", it would probably get done faster ;)
On another note, here's a debug build of modelview.
http://blueplanet.fsmods.net/E/modelview_Debug.7z
Please try this one as well.
-
Gets up to the main screen.
(http://i229.photobucket.com/albums/ee67/waternz/other/modelview-E1.jpg)
-
I'm getting the same thing as Water.
If I hit ignore once, I get this message:
(http://www.ciinet.org/kevin/myimages/Asserterror2.PNG)
-
Both builds ran flawlessly here.
I tested them using some HTL models from FSport and TVWP, with no textures. I used "rotate parts", "thrusters", "shields", and rotated the models around.
The only difference I see is that my MODVIEW32 installation is an old one that came with Turey's installer.
It's located in D:/fs_open/tools/MODVIW32
Any other useful data that I could provide?
EDIT: D'oh. I'm also running Win7 beta.
EDIT2: Yet another d'oh. 64 bits version. Sorry Talon 1024. :D
-
Any other useful data that I could provide?
What version of Windows are you running? I wonder if it has something to do with the fact that my version of Windows XP is a 32-bit OS.
-
Just edited the post, adding more data. ;)
-
Yes, I suspect me not defining a Windows version to target might be the culprit, as VS then defaults to the highest possible version (and since I'm on Win 7 x64 Pro, well....).
Please try out these versions. The first one was compiled using the VS2010 compiler, the second one with the VS2008 one. Both are compiled (I hope.... :nervous:) using the "XP" includes, so they should work....
http://blueplanet.fsmods.net/E/ModelView%20v10%20Toolset.7z
http://blueplanet.fsmods.net/E/ModelView%20v9%20Toolset.7z
-
I managed to reproduce the bug with the old version running over wine "emulating" XP.
The newer versions ("ModelView v9 Toolset" and "ModelView v10 Toolset") fixed it (at least in wine).
All versions still run flawlessly on Win 7 so far.
-
Both v9 and 10 work on xp
Two limits some models have exceeded though
Sub model with greater than 10000 points and
Docking points above 32
-
You can probably use the latest includes, just set the lowest target you want to run on. 2k if possible or at least XP.
-
Which is what I did. WinXP is defined as the lowest supported Windows version now.
-
Both v9 and 10 work on xp
Two limits some models have exceeded though
Sub model with greater than 10000 points and
Docking points above 32
Hmm. Are those craft publically available? If not, how far do I have to bump the limits?
-
Docks are now at 1000 but will probably be dynamic at some point.
-
Hmm. Are those craft publically available? If not, how far do I have to bump the limits?
Highest values I've seen so far.
#6007 Too many SUBJ chunks = 402
#6029 Too huge sub-model points = 24578
#6018 Too many docking points = 864
-
Here are two examples of models that cannot be opened because of limits:
PVI Karnak (karnak.pof) in the FSPort mediavps (mv_fsport.vp) (Error #6029: Too huge submodel, too many points defined.)
Vishnan Preserver (xdread.pof) in Blue Planet (bp-visuals1.vp) (Error #6017: Contains a thruster with too many glowpoints.)
Also, your build of MODVIEW32 cannot save mass values to FS1 models, whereas MODVIEW32 legacy can.
-
A new build for you guys to play with:
http://blueplanet.fsmods.net/E/ModelView%2011A101.7z
Several changes have been made in this:
--The nonstandard BCMenu and coolcontrolsmanager additions have been removed
--The texture loading code has been reworked in preparation for additional texture formats
--Dependence on dm_tools.dll and other assorted bits of legacy stuff has been removed.
--Registry access code has been reworked to stay the **** away from HKLM (you may have to re-set a few settings)
--Stuff related to Descent and Red Faction has been exorcised
--Minor UI changes
--A few limits have been raised
As for the Mass value issue, no idea what happened there. I haven't really touched that part yet, but since the model loading/saving code has to be reworked anyway....
-
Large submodels seem to be covered now. What is the current limit on SOBJ chunks?
-
--The texture loading code has been reworked in preparation for additional texture formats
--Dependence on dm_tools.dll and other assorted bits of legacy stuff has been removed.
--Registry access code has been reworked to stay the **** away from HKLM (you may have to re-set a few settings)
--Stuff related to Descent and Red Faction has been exorcised
--A few limits have been raised
:yes: ^^
I can't download it now, but I will do it later.
-
Right, so it seems my texture loading rewrite broke texture loading slightly. I fixed it and uploaded the fixed version. Get it at the link above, or this one here:
http://blueplanet.fsmods.net/E/ModelView%2011A101.7z
Large submodels seem to be covered now. What is the current limit on SOBJ chunks?
I set it to 200, which seemed sufficient for now.
Well, next project, after introducing support for more texture formats, will be to remove those static limits.
As for allowing collada import, that's going to be a wee bit trickier. Unless someone can point me to the source for the PCS2 collada importer, I'll have to figure out how to do the conversion stuff myself, which will take a while.
-
I'll see if I can find a copy of it on my local files, but it might be in one of the main PCS2 threads, unless it was an attachment in which case it's long been deleted. Or PM/IM Spicious.
-
I did. Haven't received an answer yet.
-
Enjoy!
[attachment deleted by admin]
-
Thanks for that! Let's see what I can learn here.
I'll also stick it on the ModelView svn, to prevent attachment rot.
EDIT: For anyone interested in this, here's how to get it to compile on Windows:
1. Get the PCS2 source
2. Download the wxWidgets SDK (http://www.wxwidgets.org/downloads/). I used version 2.8.11. Install it, and build the "DLL Debug" and "DLL Release" targets. This guide (http://wiki.wxwidgets.org/Microsoft_Visual_C%2B%2B_Guide) may come in handy.
Remember to build it with wxUSE_GLCANVAS set to 1.
3. Download the colladadom SDK (http://kent.dl.sourceforge.net/project/collada-dom/Collada%20DOM/2.1/colladadom.zip). Install it, and build both the debug and release dlls.
4. Download and install the DevIL SDK (http://openil.sourceforge.net/download.php). No building necessary, as it comes precompiled for your convenience.
5. Download and install the SDL SDK (http://www.libsdl.org/download-1.2.php)
6. Get the boost package (http://sourceforge.net/projects/boost/files/boost/1.43.0/), and install it.
7. Open the solution. Adjust directories as necessary.
8. Solution should now build.
-
Awesome work, The E. :)
-
Haven't tried saving yet but did a quick run through.
Getter G, 23k Check!
TOS CBstar, 82 turrets, Check!
And numerous other fighters I couldn't open before, Check!
Excellent work, pls keep it up!
I for one appreciate the more user-friendly functionality of MV!
-
Just loaded up my own models in ModView, petty good so far. When will .dds preview support come around. Sure would be nice to see the textures before they make it ingame.
Great job so far! Oh, and .dea support is a MUST. Otherwise I won't be able to use it period...
-
Current plan in that regard is to import PCS2's model loading/handling code. That is, turn that part of PCS2 into a standardized library that can be used by both PCS2 and ModelView. The texture loading code would improve as a side effect.
-
I'm guessing the Collada code we got from Spicious will likely be utilized in some way as well, so you can probably count on your DAE support.
-
Yeah, that's part of it. Would be a bit stupid to import all that code from PCS2 without getting the model import code as well.
-
A brand new program would be precisely the wrong thing to do; why start from scratch when you can build upon something that already exists?
The consensus seems to be that the PCS2 codebase is very hard to work with, and this is complicated by the facts that 1) the PCS2 code repository isn't available for new members; 2) the most recent PCS2 code isn't even in the repository. In contrast, the Modelview source code is freely available, well organized, and well documented.
In view of these facts, focusing on Modelview would seem to be the better choice.
1) Yes it is, how many times do i have to say "PM me for access"?
2) what is in the repo is not very out of date. HEAD is unstable, but the 2.0-fixes branch is
3) Collada still needs merged.. i think i have a changeset against HEAD around here somewhere
don't fork out of my repo FFS
[edit]
and instead of *****ing about how un-userfriendly PCS2 is how about you make suggestions on what you think needs improvement - to me (and I assume to bob who did the UI) it is no more difficult to use than Modelview - PCS2's UI was designed to copycat the basic layout of Modelview
my wife and I are playing WoW which is what is keeping me distracted from playing FS2 again and thus working on PCS2.
[edit2] what files are missing from the repo?
-
I'm sorry, what are you saying here? Are you saying I shouldn't take your code to improve ModelView? Are you saying that PCS2 should only be developed with you as project lead? If so, then you saying "I'm too busy to work on PCS2 because I'm playing WoW" is sort of giving us all the bird, isn't it?
As for PCS2's UI.....where to start. Suffice it to say it is nowhere near as user friendly or accessible as ModelView's, from my POV. ModelView gives you a lot of information about the model that PCS2 doesn't give you. PCS2 gives you error messages that are not very helpful (The "great flameing graphics error of DOOM!", for example).
Seriously, just do a side-by-side comparison of the two programs, and compare the functionality.
-
I'm sorry, what are you saying here? Are you saying I shouldn't take your code to improve ModelView? Are you saying that PCS2 should only be developed with you as project lead? If so, then you saying "I'm too busy to work on PCS2 because I'm playing WoW" is sort of giving us all the bird, isn't it?
As for PCS2's UI.....where to start. Suffice it to say it is nowhere near as user friendly or accessible as ModelView's, from my POV. ModelView gives you a lot of information about the model that PCS2 doesn't give you. PCS2 gives you error messages that are not very helpful (The "great flameing graphics error of DOOM!", for example).
Seriously, just do a side-by-side comparison of the two programs, and compare the functionality.
1) use my repo - i don't mind if you post improvements (please do) - but use my repo. I intend to move it into SVN soon as I fix a stability problem in head. I'll try to fix that this weekend and get it moved.
2) please don't rip code from my app to use in modelview. if you have a problem with the UI, improve it.
3) i'm saying keep it in my repo
4) I forgot to add the "but i'll try to set aside some time to work on it" .. it'll be easier once my wife's new char hits 80
UI
A) What information does Modelview give that PCS2 doesn't?
B) that's one of Bobboau's errors.. we can clean up them
C) side by side comparison of functionality? PCS2 wins hands down as it is designed to be a model converter and meta data editor from the start. Modelview bolted those features on as an afterthought.
-
Why will no one skip SVN in the VCS chain?
I quite like the idea of splitting the pcs2 backend from the frontend. (I was going to suggest it...)
-
Why will no one skip SVN in the VCS chain?
I quite like the idea of splitting the pcs2 backend from the frontend. (I was going to suggest it...)
A) because the alternatives are not hugely better than SVN
B) the backend and the frontend are not very tightly bound to start with. the changes in head that made it unstable are actually UI changes
oh.. and i really need to go.. you've made me late for work!
-
why are we having this fight? having the two programs being developed will be an asset to the comunity, at the end of the day one program will always be better at some things and the other better others, this will happen because the prioritys of one dev group will be different to the others so can we drop the pissing contest, agree to share info when needed and see what happens, you are both SCP and adults so start behaving like it not two spoilt brats trying to one up your school projects
-
[edit2] what files are missing from the repo?
1>c1xx : fatal error C1083: Cannot open source file: 'wxCTreeCtrl.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'PCS2_MainPanel.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_weapons.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_turrets.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_thrusters.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_textures.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_subsystems.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_subobjects.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_paths.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_header.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_glows.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_docking.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'kaz_templates.cpp': No such file or directory
2) please don't rip code from my app to use in modelview. if you have a problem with the UI, improve it.
Could you elaborate on why I shouldn't? What I am doing pretty much amounts to ripping out the backend of modelview and replacing it with PCS2's. Given the limited amount of coders available, that seems to be the best option, as having a common backend for both allows shared development.
C) side by side comparison of functionality? PCS2 wins hands down as it is designed to be a model converter and meta data editor from the start. Modelview bolted those features on as an afterthought.
Okay. One thing ModelView does better is support for files in VPs. Its Textures list is better as well (as in, it gives you a preview of the texture separated from the model).
It's various toolbars are better organized (IMHO). It has a working shield editor.
To be fair, PCS2 is better in that it has editors for all model information. ModelView kinda lacks editors for dockpoints, paths and glowpoints at the moment, which is just a bit embarassing.
Now, UI design is something where you can't please everybody. For myself, ModelView as a whole presents a friendlier UI that I, for one, find easier to work with.
why are we having this fight? having the two programs being developed will be an asset to the comunity, at the end of the day one program will always be better at some things and the other better others, this will happen because the prioritys of one dev group will be different to the others so can we drop the pissing contest, agree to share info when needed and see what happens, you are both SCP and adults so start behaving like it not two spoilt brats trying to one up your school projects
As far as I am concerned, this is not a fight. Kazan has his opinions on the subject of developing PCS2, I have mine. Other people have other opinions. For the moment, I am just trying to understand the reasoning behind Kazan's opinions. But I don't react particularly well to being told that I can't or shouldn't do something without some reason.
-
A) because the alternatives are not hugely better than SVN
This might be relevant if the repo was already using SVN. Given that it isn't, the fact that the alternatives are in fact better than SVN makes them the obvious choice.
-
As a recent convert to git, I would have to support that statement.
-
Ah crap, I think I continued this discussion in two other threads before I realized this is where it's really going on.
Git please. For god's sake please move it to git. SVN is not a serious improvement over CVS, whereas git (or a proper DVCS in general) is orders of magnitude better than either CVS or SVN. And the fact that Sourceforge has Git, Mercurial, and Bazaar support should mean it's not a problem at all to keep it on SF.
Kaz, I would love to see PCS2's interface developed over ModelView, but no level of improvements are going to convince everyone that it's hands down more usable. We can probably win some over with some fixes and changes, but I don't see any reason to stifle the recent interest in modernizing ModelView, if it at the same time will bring improvements to PCS2. I've already discussed this with The_E on IRC, and I think that using a common backend for bother apps would benefit everyone in the end. If you want, it doesn't even need to be in the core, the beauty of Git means that The_E could work on the PCS2 core backend and keep the ModelView stuff in a separate branch, and they could still easily stay in sync, much much more easily than with Subversion. Breaking the PCS2 backend into a fully discrete library/object would probably be very helpful though.
The_E, Kaz is right in that the reason ModelView is lacking some of those 'embarrassing' features is that they were never planned. The editing capabilities were a bolt-on afterthought as he said. PCS2 was designed from the start as an editor, and even went through a long design process where any suggestions to the interface could have been suggested, but what we have now is what it evolved into. If it's less than desirable, that is as much the modding community's fault as theirs :P
But seriously guys, there's no reason we can't all pull together into one active project where everyone wins, is there?
-
The_E, Kaz is right in that the reason ModelView is lacking some of those 'embarrassing' features is that they were never planned. The editing capabilities were a bolt-on afterthought as he said. PCS2 was designed from the start as an editor, and even went through a long design process where any suggestions to the interface could have been suggested, but what we have now is what it evolved into. If it's less than desirable, that is as much the modding community's fault as theirs
Well, I like to think that the reason these things are missing has more to do with the project being abandoned 8 or so years ago. I mean, it's clear that an implementation was planned, but not started. Should be easy to add though, once I get the backend sorted.
Agreed on all other points, though.
-
Abandoned or not, they weren't part of the original plan. I'm downright amazed that ModelView works as an editor half as well as it does.
-
me personally i wouldnt mind seeing as many of the classic modeling tools brought up to spec. modelview, styxx's max converter (especially this one), maybe even aurora (or maybe not). pcs2 as an editor is pretty good now if you disregard the ui related crashes. my holy grail of model conversion however has always been to convert a model without any post conversion editing, a quick test of the collada importer kinda saw that happen (how comprehensive it is has yet to be seen, by me at least). my biggest problem with pcs2 is that it hasnt changed a whole lot lately, the collada builds are still kinda buggy. i had even considered trying to work on the code. at least a couple times in the last 4 months i attempted to compile it with very little success (should give the_e's procedure a go next time i feel up to coding). id rather use svn simply on the grounds that its what fs2 uses and i could just use one client for everything. i had a problem recently with the icon overlays in tortoise svn because i had also installed tortoise cvs so i could download pcs2.
-
[edit2] what files are missing from the repo?
1>c1xx : fatal error C1083: Cannot open source file: 'wxCTreeCtrl.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'PCS2_MainPanel.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_weapons.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_turrets.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_thrusters.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_textures.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_subsystems.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_subobjects.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_paths.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_header.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_glows.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'pe_docking.cpp': No such file or directory
1>c1xx : fatal error C1083: Cannot open source file: 'kaz_templates.cpp': No such file or directory
ah.. compiling on non-windows... if you're using the Makefile it's out of date... use SCons for non-windows builds
2) please don't rip code from my app to use in modelview. if you have a problem with the UI, improve it.
Could you elaborate on why I shouldn't? What I am doing pretty much amounts to ripping out the backend of modelview and replacing it with PCS2's. Given the limited amount of coders available, that seems to be the best option, as having a common backend for both allows shared development.
for starters - because it's my code and I asked you not to. because:
A) Modelview's frontend is non-portable
B) modelview's frontend wasn't designed to be an editor from the start
C) because I'd rather concentrate programming effort on something worthwhile
C) side by side comparison of functionality? PCS2 wins hands down as it is designed to be a model converter and meta data editor from the start. Modelview bolted those features on as an afterthought.
Okay. One thing ModelView does better is support for files in VPs. Its Textures list is better as well (as in, it gives you a preview of the texture separated from the model).
It's various toolbars are better organized (IMHO). It has a working shield editor.
To be fair, PCS2 is better in that it has editors for all model information. ModelView kinda lacks editors for dockpoints, paths and glowpoints at the moment, which is just a bit embarassing.
Shields are mesh, and should be edited as such. Implementing mesh editing was on the TO DO list for PCS2. What's wrong with PCS2s support for textures in VPs? and i'm not sure what you mean "preview of the texture separated from the model"
Now, UI design is something where you can't please everybody. For myself, ModelView as a whole presents a friendlier UI that I, for one, find easier to work with.
but what is substantially different about the UI that makes this so, that it isn't just easier to correct in the already cross platform portable UI code in PCS2?
As far as I am concerned, this is not a fight. Kazan has his opinions on the subject of developing PCS2, I have mine. Other people have other opinions. For the moment, I am just trying to understand the reasoning behind Kazan's opinions. But I don't react particularly well to being told that I can't or shouldn't do something without some reason.
I asked you nicely not to fracture my codebase, and not to rip my code for use in other applications.
if it comes down to it: PCS2 isn't GPL. In fact I haven't applied a specific license to it. any and all usage of my code goes through me - it's my code.
-
oh.. and as i speak i'm cleaning up my copies of the source for prepping to migrate to (by popular demand *Drumroll*) SF's Git
Spici would you PLEASE PM me your SF.net user so that I can add you to write access, and would you please merge all your changes into 2_0-fixes - i'm going to base the new Repo off stable (2_0-fixes)
[edit]
actual branch name is "stable_2_0_fixes"
i might merge head.. i'm not sure.. i'm goin through the diffs between stable_2_0_fixes now.. one of them was reverting out my templates (kaz_*) because of updating to modern MSVC copies with decent STL support
[edit2]
yes.. commit your changes to HEAD please... there is more development in it than i thought. some of the changes have made the codebase unstable but i'll try to debug them out before i migrate us to GIT - but i need your changes first Spici
-
for starters - because it's my code and I asked you not to. because:
A) Modelview's frontend is non-portable
B) modelview's frontend wasn't designed to be an editor from the start
C) because I'd rather concentrate programming effort on something worthwhile
A and B are reasons why the conversion will be nontrivial (A fact of which I am aware).
C is questionable. There are people out there who prefer modelview to PCS2. How a combined effort to bring both applications on the same level to serve both groups is not worthwhile is quite beyond me.
And while you are fully correct in saying that it is your code, you also made the decision to make that code public. That means that everyone, even one such as I, can go and take it to do whatever I want with it, in accordance with the license under which that code was released.
And I am quite sorry, but in this instance, I am not inclined to bow to your wishes just because you ask me to. If there are technical reasons why something is impossible, I am willing to listen to them. If it boils down to "because I don't want you to", I will ignore you. That's how I see it.
ah.. compiling on non-windows... if you're using the Makefile it's out of date... use SCons for non-windows builds
Actually, no. I'm using Visual Studio 2010.
Shields are mesh, and should be edited as such. Implementing mesh editing was on the TO DO list for PCS2. What's wrong with PCS2s support for textures in VPs? and i'm not sure what you mean "preview of the texture separated from the model"
On the TO DO list. Good. Nice to hear.
As for the support for textures in VPs, nothing is wrong with it per se. But if you look at modelview, it allows you to look at models in vps without extracting them from the vp first. It also allows you to edit those models (although, and this is certainly a good thing, it can't repackage those vps).
but what is substantially different about the UI that makes this so, that it isn't just easier to correct in the already cross platform portable UI code in PCS2?
For me personally, the fact that I can use the standard MFC editors on it is a big plus.
I asked you nicely not to fracture my codebase, and not to rip my code for use in other applications.
if it comes down to it: PCS2 isn't GPL. In fact I haven't applied a specific license to it. any and all usage of my code goes through me - it's my code.
Why would my efforts fracture your codebase? Am I committing breaking changes to your version control system without consulting you? Am I distributing builds that are not approved by you? Am I asking you to fix my bugs?
Also, since your code is hosted on SourceForge, permission to fork your code is sort of implicated. Unless you decide to make PCS2 closed source, in which case already released code is still released, and the fork can continue anyway.
I am very sorry about this escalation, but I have little to no respect for this sort of behaviour. What I see here is your desire to keep the development effort centralized. Which I can understand and appreciate. What I cannot understand is your insistence that ALL development has to be approved by you.
-
[redacted]
discussed on IRC
The_E (And anyone who wants to help) is going to work on enhancing PCS2's UI To be much more awesome
I'll try to work on the back end as time allows to start implementing my backlog of 2.1 roadmap changes
anyone else who wants to help feel free.
now... SF.net GIT, or Github? which is better to people. i'm looking through OSI-approved licenses tryin to decide which one to officially put the code under.
-
I think we have reached an agreement here (well, not here, but on #hard-light, if you get my drift).
Bottom line is that I will probably discontinue working on modelview, but that PCS2 will get a much better UI.
As for project hosting, I'm fine with either of those. Although I prefer github for its cleaner layout.
-
I think we have reached an agreement here (well, not here, but on #hard-light, if you get my drift).
PCS2 will get a much better UI.
this makes me a very happy coder :D
[edit]
The E got me a copy of Spici's changes.. and i've merged them.. so i should be able to move us onto Git this weekend
-
Excellent. Kaz, if you'd like to set me up with access as well my SF username is also chief1983. And if anyone needs any Git help, I'm available. I mostly work with command line, but I might be able to figure out TortoiseGit if anyone needs help with that. But command line Git is so powerful. Bob, you can ditch TortoiseCVS and just use TortoiseSVN if you want.
One of the nicest things about Git is that if we wanted to move off of Sourceforge later, it's painfully easy compared to Subversion. So I'm not really picky at all about that either, put it wherever you want.
-
And so it came to pass that the great coding forces of Kazan and The_E had been in conflict. They each possessed their own visions of what the community needed and desired. Then, just when it seemed that the gulf between them would widen to insurmountable proportions, stunting further progress, the two came to see the reason in each others viewpoints and concluded that for the good of the community, that they should join forces. Combining their individual skills, Kazan and The_E, along with their supporters, would together strive to do that which neither could do on their own... create the most stable, feature-filled, and user-friendly model editing tool in existence, and in doing so, usher in an era of modding prosperity for all in the Freespace Community.
Here's to the new alliance. :yes:
No offense to the other team members whose names I left out... it just seemed to fit.
-
Yeah I don't know what happened in IRC, but I'm now extremely happy. I have _got_ to go read Anna's logs...
-
I am skeptical. I would much rather see stuff moved into ModelView from PCS2 than the other way around.
Also...
if it comes down to it: PCS2 isn't GPL. In fact I haven't applied a specific license to it.
The attachment disagrees with you. (See below.)
any and all usage of my code goes through me - it's my code.
We had this argument back when you first started adding FS2NetD to FSO. After Inquisitor confronted you, you agreed to not restrict the rights of others to play with code you had publicly released as open-source. Even though this is not FSO code, the same principle applies.
I don't know the details of the agreement between you and The E (which I plan to inquire about) but I felt these points should be mentioned.
[attachment deleted by admin]
-
Goob we've already come to an agreement. kindly redact your post.
PS: grandfathered in at SF
PPS: courts would most likely find that clause of their TOS an unconscionable clause
PPPS: that agreement with Inqui only applied to code I added to FSO.
-
Please take this to PMs.
-
I am skeptical. I would much rather see stuff moved into ModelView from PCS2 than the other way around.
yeah.. modelview with its hammered **** gui code (gui might look pretty.. buts its code isn't) that is non-portable, and wasn't designed to be an editor from the ground up.
we've already settled this. The_E is going to make an awesome GUI for PCS2.
if it comes down to it: PCS2 isn't GPL. In fact I haven't applied a specific license to it.
it is now.. so the rest of your post is pointless
-
Registered on sf as spicious.
I look forward to the awesome (and hopefully cross-platform) GUI.
-
added
-
I don't suppose a chatlog of the discussion could be posted, I'd like to know what changes are being planned. I just read this whole ****ing topic to find out what changes are planned only to find 'we talked about it in chat'. hearing 'redoing' the UI is making me a little nervous, because the UI was designed to be easily extensible and dynamic. we aren't going to end up with having to manually read and write to the control for everything are we? were not going to have controls that stay the same shape and size regardless of the parent panel they are in are we?
now as far a what was being discussed earlier about using the PCS2 backed and the modelview frontend, I actually think this might be a worthwhile project, as it would involve decoupling the POF classes from the UI and allow it to be easily integrated into other projects. this would necessitate breaking PCS2 in 2 though a POF library and PCS2 proper, which again I think would be a good idea.
-
Supposedly there isn't too much coupling already, as before this plan was set in motion The_E believed he had already isolated the PCS2 backend into its own library. But yes, I've been pushing for discrete objects and libs for lots of things, way back to the PCS2 Collada support. It would have allowed various front ends to make use of the Collada conversion code without being dependent on PCS2 or changes to its interfaces.
-
making it possible to run pcs2 in a command line mode would be pretty cool for anyone who wants to write custom scripts for conversion. all editing should be possible with this meathod. say if i wanted to make a large detailboxed terrain map from a large texture and topographic map. i could write a script that would automatically cut up the texture, generate geometry for them at various levels of complexity, and feed all the neccisary model and texture data into pcs2, setting up all the detail box subobject flags for each tile automatically, then save to pof. im sure there are other things like that wich could be done with such an interface.
-
Yup, having a separate lib for all that stuff would make that far more trivial. But again, it sounds like it's pretty close to that already, so it might not be too hard to do.
-
yes the backend is trivial to isolate from the frontend. it was designed that way intentionally.
no to modelview frontend.
-
if someone else wants to do it and it helps improve modularity of pcs2 / pof code why not? for an open source proponent you are awfully clingy with what happens with your code outside of your control.
-
Modelview's UI is written in MFC, i really don't want any code i wrote behind an MFC application. I cannot stop you, but i really would rather it not be done.
-
There's definitely other purposes it could serve though too. Nuke's suggestion is a good one. Unit testing too.
-
by that logic a billion open source libraries wouldn't have been written because of fear that someone would use them in a ****ty project.