Matrix Games Forums

Australian Design Group (ADG) is . To the best of my knowledge they have done the following:

• They obviously are the original author of the tabletop game.
• Provided the original CWiF code to be reworked.
• Provided access to the original author of CWiF, Chris Marinacci (my apologies if mispelled).
• Harry Rowland has clarified numerous rules questions necessary for programming the game.
• Harry Rowland has given his blessing on MWiF rules deviations where they are necessary. Thereby keeping it inline with his vision of WiF

The former items have been stated by ADG, Matrix or others connected with the project in various ways, e.g. Patrice Forno.

Who knows what else they have supplied behind the scenes or what may be in their contract(s) with Matrix and/or Steve.

The formatting for this is messed up. I am redoing it.

While we wait for the next update from Steve, I thought I'd take the opportunity to post a summary of Steve's monthly postings. Hats off to him for providing all this information as he has been working on the game! It is rare to see that (actually, I don't think I've ever seen it).
I think this summary shows how hard is it to estimate a project of this size:

Sorta makes you wonder if that spring 2006 prediction was made with a straight face or with beer shooting out of his nose.

ORIGINAL: pakron

While we wait for the next update from Steve, I thought I'd take the opportunity to post a summary of Steve's monthly postings. Hats off to him for providing all this information as he has been working on the game! It is rare to see that (actually, I don't think I've ever seen it).
I think this summary shows how hard is it to estimate a project of this size:

Code: Select all

 Date            Estimated Release       Delay           Reason for delay                                Quotes from Steve 
 Aug 2005        Spring 2006                                                                             “Shooting for early spring 2006”
 Sep 2005
 Oct 2005        Spring 2006                                                                             “Spring 2006 not hopeless yet”
 Nov 2005        Spring 2006                                                                             “Spring 2006 not hopeless yet”
 Dec 2005                                                Graphics delayed 3 months                       “Hopes for the target date of spring 2006 fade.”
 Jan 2006        Summer 2006             +3 months                                                       “Best guess”
 Feb 2006        Summer 2006                                                                             “Best guess”
 Mar 2006        Summer 2006                                                                             “Best guess”
 Apr 2006        Summer 2006                                                                             “Best guess”
 May 2006        Late Sep 2006           +2 months       No actual delay according to Steve              “My definition of summer is not in common usage”
 Jun 2006        Dec 2006                +3 months       Graphics won’t be completed in time
 Jul 2006        Dec 2006                                                                                “still my best estimate”
 Aug 2006        ?                                       Lost month of July                              “December 2006 is too optimistic”
 Sep 2006        1st Q 2007              +3 months
 Oct 2006        1st or 2nd Q 2008       +3 months
 Nov 2006        ?
 Dec 2006        July 2008               +4 months                                                       “I did my first detailed project plan in mid-October and have more or less kept on track against that to date”
 Jan 2007        July 2008
 Feb 2007        July 2008                                                                               “still on track with the schedule from last October”
 Mar 2007        July 2008										“I have lost about 1 week”
 Apr 2007        July 2008                                                                               “I lost another week”
 May 2007        July 2008
 Jun 2007        Origins 2008 (June 2008)
 Jul 2007        Origins 2008 (June 2008)								
 Aug 2007        Origins 2008 (June 2008)
 Sep 2007        Origins 2008 (June 2008)
 Oct 2007        Origins 2008 (June 2008)
 Nov 2007        Origins 2008 (June 2008)
 Dec 2007        Origins 2008 (June 2008)                                                                “To accomplish this [date] I had to move a few features”
 Jan 2008        Origins 2008 (June 2008)                                                                “more or less stayed on track to deliver MWIF product 1 for Origins 2008”
 Feb 2008        Dec 2008                +6 months       Bugs                                            “What consumes hundreds of hours of my time are inexplicable bugs.”
 Mar 2008        Dec 2008                                                                                “I am still planning on attending Origins 2008 and demonstrating MWIF”
 Apr 2008        Dec 2008
 May 2008        Dec 2008
 Jun 2008        Dec 2008                                                                                "David Heath (Matrix Games) has persuaded me to not attend Origins 2008”
 Jul 2008        Dec 2008
 Aug 2008                                                                                                “There are too many bugs being found by the beta testers and the work on NetPlay has yet to reach the testing stage.”
 Sep 2008        Late March 2009 	+3 months                                                       “There is still some uncertainty about that date, mostly related to NetPlay.”
 Oct 2008        Late March 2009                                                                         “If Dan can get it working in October, then we should make the above date.” 
 Nov 2008        Late April 2009 	+1 month        Dan was unable to find time to work on NetPlay
 Dec 2008        Late April 2009
 Jan 2009        Spring 2009             +1 month
 Feb 2009        July 27, 2009           +2 months                                                       “Firm release date “ , “we agreed on July 27th as the release date. I am comfortable with that.” , “some slack in the schedule”
 Mar 2009        July 27, 2009
 Apr 2009        July 27, 2009
 May 2009        July 27, 2009										“I am gaining ground on the beta testers but it is measured in inches”
 Jun 2009        October, 2009           +3 months       Bugs and other tasks                            “Too many bugs and too many other tasks remain”
 Jul 2009        November, 2009	+1 month
 Aug 2009        First week of November, 2009
 Sep 2009                                                                                                "I really don’t know how long it will take me to fix all the bugs I am seeing”
 Oct 2009                                                                                                “I still don’t know how long it will take me to fix all the bugs.”
 Nov 2009                                                                                                “I still don’t know how long it will take me to fix all the bugs, but I am making reasonable progresson them.”
 Dec 2009                                                                                                “I still don’t know how long it will take me to fix all the bugs, but I continue to make progress on them.”
 Jan 2010                                                                                                “Tasks for 2010 Finish MWIF product 1 so I can buy a large screen, flat panel, HD TV.”
 Feb 2010        											“I spent virtually all of January trying to get Delphi 2010 and Theme Engine 9.10 working correctly.”
 

Nice.[:)]

===
Things I didn't expect when I signed my contract in July of 2005:

1 - that the map and unit graphics would require ~9 months of my personal time. Patrice also did hundreds of hours of work on the map. I had thought that the graphics artist would deliver unit and map bitmaps which I could use directly, which was not the case. And the last major delivery of the bitmaps from the graphics artist didn't arrive until in December of 2007.

2 - that the CWIF forms could be converted automatically using Theme Engine. While that was partially true, in fact, the CWIF forms were small in size, used small fonts, button sizes, and unit depictions (zoom level 4). I needed to go through all the ~105 forms from CWIF and modify them substantially. There were several very complex forms which I reworked extensively: Start New Game, Setup, Scrap, Air-to-air Combat, Land Combat, Production Planning, ... and others. I also ended up adding 50+ new forms, some of which were fairly elaborate: Lend Lease Air Units, Victory Hexes, Flyouts, Naval Review Details & Summary, and others. I created new unit graphics of zoom levels 5, 6, and 8 for use in the forms so the unit depictions on the forms were more legible; and then I swapped out the zoom level 4's with 5's and 6's. The zoom level 8's provided maximum visibility on individual units (e.g., when breaking down a corps).

3 - that the rules from WIF FE would be all that I would need, and the CWIF code implemented them. In actuality, the WIF FE rules were only a starting point, with hundreds of modifications, corrections, and clarifications necessary before I could edit RAW into RAC (Rules as Coded). RAC serves as my design document on how MWIF implements the WIF FE rules. CWIF was not up-to-date with the changes to the rules from 2003. Additiomally, ADG keeps tweaking the rules, which in general is good - just not for the programmer. I finally set a cut-off date of July 2008 for rules changes (there have been some by ADG since then, but for the most part, they are not part of MWIF). I had to review all the code that implemented rules to ascertain whether or not they followed RAC correctly.

4 - that zero work had been done on the Players Manual, and that creating it would be a lot of work. I had thought that RAC would be enough of a "Player's Manual", but that was delusional on my part. New players to WIF do not want to sit down and read the 150+ pages of RAC before playing the game. The need to provide help to new players resulted in the creation of the 120+ pages of the Picture and Text Tutorials, the 13 Training Videos (9 completed so far), 300+ pages of help messages embedded in the game (for 150+ forms and 150+ phases/subphases), and the Plpayer's Manual, which whenfinalized will run to ~300 pages. I still need to create the 10 Interactive Tutorials, though those will follow the Training Videos.

5 - that I would get programming help on coding NetPlay. I ended up doing that myself.

6 - that the CWIF code would be more structured than it was. In CWIF the Main routine was 23,000 lines long (I have since reduced it to 8,500). Without any documentation on the 100,000 lines of CWIF code, and some of its less than desireable programming practices, I spent months of time adding comments and revising the code so it is more structured. The improved structure was essential for adding PBEM, NetPlay, and AI.


I could go on, but those were the big surprises.[X(]

This is a duplicate of the report in the thread Monthly Reports. Unlike that thread, this (the When thread) is open to comments.


April 2, 2010 Status Report for Matrix Games’ MWIF Forum


Accomplishments of March 2010

Project Management
I monitored all the threads in the MWIF World in Flames forum daily.

Hardware and Software
I completed replacing all the assembler routines with equivalent Pascal routines. Doing so also eliminated any code that modified the call stack. I still have glitches when Delphi 2010 IDE (Interactive Development Environment) executes MWIF in debug mode, but it continues to improve (or at least I have that impression). Once the Assembler mysteries were gone I was able to focus on Theme Engine bugs.

Theme Engine 9.10 caused me some new grief. One of its new features is the ability to animate buttons (TTeButton) so they can be programmed to blink and do other stuff. However, the cost of this change is 5 more bitmaps for each TTeButton. MWIF can not afford that - and I have zero interest in animating buttons. The program is already close to tapping out the available bitmaps under Windows XP and given that MWIF has 750+ TTeButtons (i.e., 3750+ new bitmaps), this change in Theme Engine caused the program to fail do to insufficient “Windows resources”. So, I replaced all the TTeButtons with TTeSpeedButtons (which do not have the sexy animation feature).

Although that global change was done with one command and executed in 2-3 seconds, there were other code changes required. The compiler caught most of those and I was able to get a clean compile and execution within a half hour. Annoyingly, the differences were more extensive. A TTeButton has embedded defaults which were used extensively within MWIF. A TTeSpeedButton has no defaults so I had to explicitly code stuff that had been done automatically previously. For example, when you click on an OK button (and it is a TTeButton), Windows automatically closes the form, and records that you pressed ‘OK’. It does something similar for Cancel, Yes, and No buttons. Basically, I spent 3-4 days making the full conversion to TTeSpeedButtons.

After all these attempts to remove suspects, I was still getting weird errors. Therefore I disabled the execution of the Theme Engine themes. At that point I was able to debug in a more normal manner, and I have cleaned up what I hope to be all the newly created bugs do to the conversion from Delphi 2007 running under Windows XP to Delphi 2010 running under Windows 7. I even fixed some older bugs I encountered along the way.

Beta Testing
I released version 4.01.00 to the beta testers last week. I expect to release 4.01.01 later today and return to meeting my goal of releasing a new version every 5 days or so. I would like to do 6 new versions/month. Once things settle down and I have restored my shattered confidence in the MWIF code, I’ll reenable Theme Engine themes and see how that goes. By the way, I did spend a little over a week this month looking into modifying the Theme Engine source code (i.e., the definition of TTeButton). I had mixed success with that and decided to just go back to the original TE source code. TE consists of 100,000+ lines of undocumented code, with multiple levels of abstraction, so messing around with it is very unpleasant.

Saved Games
Fixed a bug with recording the movement of factories. Now saved games record which factories have moved where and any damage factoreis have sufferred due to strategic bombing.

Map and Units
Added some more naval unit writeups from Rob/Warspite.

Scenarios and Optional Rules
Nothing new.

MWIF Game Engine and CWIF Conversion
Substantially modified/reduced the CWIF library routines by eliminating assembler routines and unused definitions of fundamental components (i.e., hash tables).

Player Interface
Fixed some bugs in the Victory form.

Internet - NetPlay
Nothing new.

PBEM
Nothing new.

Artificial Intelligence (AI)
Nothing new.

Player’s Manual
Nothing new.

Tutorials, Training Videos, and Context Sensitive Help
Nothing new.

Historical Video, Music, and Sound Effects
Nothing new.

Marketing
Andy Johnson has returned to working on the MWIF web site (the restrictions placed on him by his employer no longer apply). Presently the web site is being reviewed by Matrix Games and the beta testers. When it will be made available to the public at large depends on Matrix Games management and other demands on Andy’s time.

Communications
Nothing new.

GG, keep plowing ahead.

"At that point I was able to debug in a more normal manner, and I have cleaned up what I hope to be all the newly created bugs do to the conversion from Delphi 2007 running under Windows XP to Delphi 2010 running under Windows 7. I even fixed some older bugs I encountered along the way."

I presume that the program/game will still run under XP when completed, correct?

Chugga-Chug like a freight train. It may not move fast and it may have a long way to go but as long as it's moving, logic dictates that the required number of precious lines of code to run MWIF, each with a precise sequence and placement, have a finite number somewhere. Thus continued progress will lead to our destination eventually.

Steve, since I find it hard to follow a lot of the terminology and concepts -I noticed on a couple reports, you grade your own progress. Would it be possible to do that more often. And after all, your assessment has got to be better than mine.

ORIGINAL: Taxman66

"At that point I was able to debug in a more normal manner, and I have cleaned up what I hope to be all the newly created bugs do to the conversion from Delphi 2007 running under Windows XP to Delphi 2010 running under Windows 7. I even fixed some older bugs I encountered along the way."

I presume that the program/game will still run under XP when completed, correct?

Win XP is the primary target market, since XP programs can be run under Vista and Win 7 using emulation. I would like to have MWIF run under all 3 operating systems without using emulation - if possible.

ORIGINAL: macgregor

Chugga-Chug like a freight train. It may not move fast and it may have a long way to go but as long as it's moving, logic dictates that the required number of precious lines of code to run MWIF, each with a precise sequence and placement, have a finite number somewhere. Thus continued progress will lead to our destination eventually.

Steve, since I find it hard to follow a lot of the terminology and concepts -I noticed on a couple reports, you grade your own progress. Would it be possible to do that more often. And after all, your assessment has got to be better than mine.

That's hard to do, since I am not sure what criteria to use. On the other hand, I can grade the performance of the software packages I am using readily.[;)] Let me be kind and just say that the JVL software library rates an A+, while the others don't.

I wonder if historical estimating would have been more accurate here. It took one reasonably capable programmer, Chris Marinacci, how many years to implement what, 40% of the MWIF feature set (no AI). A historical estimate would have looked at the only data point -- 1 FTE for 5 years = 0.4 functionality -- and asked, why should 1 FTE be able to implement 100% functionality in < 5 years?

GAO studies of NASA, etc. have repeatedly shown that high-level historical estimating -- "our last 10 satellite missions all came in between 150 and 400% over cost and schedule" -- is pretty robust.

Fred

ORIGINAL: wfzimmerman

I wonder if historical estimating would have been more accurate here. It took one reasonably capable programmer, Chris Marinacci, how many years to implement what, 40% of the MWIF feature set (no AI). A historical estimate would have looked at the only data point -- 1 FTE for 5 years = 0.4 functionality -- and asked, why should 1 FTE be able to implement 100% functionality in < 5 years?

GAO studies of NASA, etc. have repeatedly shown that high-level historical estimating -- "our last 10 satellite missions all came in between 150 and 400% over cost and schedule" -- is pretty robust.

Fred

Chris first created a program that enabled the creation of the map and units - a map and unit editor, if you will. He then built CWIF on top of that. One of the first things I did with the source code was to remove all the conditional compilation commands that enabled compiling the source code to either generate the CWIF 'Editor' or the CWIF 'Game'. I explicitly had my contract with Matrix exclude a "WIF Design Kit:.

My point is that Chris spent a lot of time not working on the 'game' per se, but rather he worked on tools to create (and modify) the game data.

Grat idea the separate post (duplicated)!

Chris first created a program that enabled the creation of the map and units - a map and unit editor, if you will. He then built CWIF on top of that. One of the first things I did with the source code was to remove all the conditional compilation commands that enabled compiling the source code to either generate the CWIF 'Editor' or the CWIF 'Game'. I explicitly had my contract with Matrix exclude a "WIF Design Kit:.

My point is that Chris spent a lot of time not working on the 'game' per se, but rather he worked on tools to create (and modify) the game data.

[&:] Why the animosity for a WiF modifier Steve? Is it that anathemic to having the game run smooth?

ORIGINAL: macgregor

Chris first created a program that enabled the creation of the map and units - a map and unit editor, if you will. He then built CWIF on top of that. One of the first things I did with the source code was to remove all the conditional compilation commands that enabled compiling the source code to either generate the CWIF 'Editor' or the CWIF 'Game'. I explicitly had my contract with Matrix exclude a "WIF Design Kit:.

My point is that Chris spent a lot of time not working on the 'game' per se, but rather he worked on tools to create (and modify) the game data.

[&:] Why the animosity for a WiF modifier Steve? Is it that anathemic to having the game run smooth?

Every change has multiple ramifications. A WIF design kit would have to be able to implement all the consequences of any change besides the change itself. This becomes an task with exponential explosion as each new change is incorporated. For example, if change A affects 4 other rules and change B affects 5, then all the interactions of A and B also have to be considered (4*5=20). Writing general purpose code that handles all possibilities is insanely complex.

This has been demonstrated vividly by ADG over the past 25 years where rule modifications upon rule modifications frequently create a gap where which rule take precedence, or the order in which the rules are to be applied is ambiguous.

ORIGINAL: wfzimmerman

I wonder if historical estimating would have been more accurate here. It took one reasonably capable programmer, Chris Marinacci, how many years to implement what, 40% of the MWIF feature set (no AI). A historical estimate would have looked at the only data point -- 1 FTE for 5 years = 0.4 functionality -- and asked, why should 1 FTE be able to implement 100% functionality in < 5 years?

GAO studies of NASA, etc. have repeatedly shown that high-level historical estimating -- "our last 10 satellite missions all came in between 150 and 400% over cost and schedule" -- is pretty robust.

Fred

Incidentally, I love NASA. there are two primary reasons for cost creep in government megatech projects: 1) requirements creep and 2) technical risk (scientists and admirals always ask for cutting edge stuff). Steve has done a good job of limiting reqts creep, but the nature of this -- probably the most complex hex to computer transform ever attempted -- is high risk.

Every change has multiple ramifications. A WIF design kit would have to be able to implement all the consequences of any change besides the change itself. This becomes an task with exponential explosion as each new change is incorporated. For example, if change A affects 4 other rules and change B affects 5, then all the interactions of A and B also have to be considered (4*5=20). Writing general purpose code that handles all possibilities is insanely complex.

This has been demonstrated vividly by ADG over the past 25 years where rule modifications upon rule modifications frequently create a gap where which rule take precedence, or the order in which the rules are to be applied is ambiguous.

_____________________________

Steve

Fair enough. Thanks for the explanation. It's certainly not worthy of holding up the release.

5 - that I would get programming help on coding NetPlay. I ended up doing that myself.

Steve reckon himself that he expected more people on the project. Hourra, that is a proof that more people on the project would have been better despite everything that some people try to argument.

Yes, doing specific side-tasks, not the core programming.