Darwinbots Forum
Code center => Suggestions => Topic started by: MrMound on October 20, 2005, 06:36:35 PM
-
one thing thats been bugging me is that I can't run a sim for very long so my suggestion is to make a version of db that works 100%(or very nearly 100%)or the time before we add more things(and by things I mean stuff like electricity) to the program. lets just get a super reliable program before we add more fun stuff.
-
OOh, sounds good. :boing:
The thing is, programming is hard, and bug fixing is boring.
Thus bug fixing is hard and boring. Now if I had a team of programmers at my disposal...
-
I would help but I don't have a compile and I don't know how to program in vb.(I don't really know how to program in anything actually)
-
Yeah...
Another problem is once you know how to code works, your less likely to go trying something that would break it, so you're less likely to find bugs than real users.
-
I was kind of thinking of doing something like that, once I learn enough about coding. Just take a fairly stable version and de-bug the heck out of it. It'd be interesting to see what a super-stable version could acheive, since it could be left on forever.
Another problem is once you know how to code works, your less likely to go trying something that would break it, so you're less likely to find bugs than real users.
I don't think I'd do this. I have a knack for thinking of oddball stuff to try. :)
-
The thing about visual basic is that it's really really really easy to pick up. You should definately give it a look some time.
-
yeah I could learn it but arn't compilers pretty expensive?
-
gentlemen ...
(I use the term loosely) ;) lol
my observations:
the possibilities for DarwinBots are great.
2.21 is a nice stable platform that works.
of course, it is somewhat limited and missing some useful features ...
but still ... the most usable.
2.36 is pretty good ... fairly stable ...
and the version I use the most ...
but with a few kinks and lacking on some features.
2.37 address some of these ... but is semi-stable at best ...
and has it's own problems and shortcomings ...
l[like run errors up the ying-yang]
2.4 has only increased the number of errors and instability.
my friends ...
it is my opinion that unless you first optimize one version ...
get it to actually work for everyone ... to address and FIX the bugs
that already exist ... actually get it to be stable and predictible ...
then adding more features and alterations to everything under the sun
will only result in more bugs, more errors, and greater instability.
if you continue to increase the number of variables ...
without addressing the problems that already exist ...
I'm afraid you are going in the wrong direction.
please, please, please consider fixing what is WRONG ...
before complicating things by adding more errors/problems.
the bottom line ...
if people can't get it to run and give them consistent results ...
they are going to give up on the program.
as much as I would love to have this work ...
this is where I find myself at the present.
I'm spending my time spinning my wheels and getting nowhere.
I only wish I knew enough about vb to actually help you out
rather than just snivel and whine ... ;)
but I don't.
so it goes.
not yet anyway.
I do have vb now ...
and will download the sources and maybe have time this winter
to play around and see if I can get into it.
is vb and the source files all I need?
or do I have to also somehow compile?
what do I need?
right now for me ...
I don't see much point in trying to run this as a serious alife simulation
the way it is going now.
anything I get to work/evolve in this environment ...
isn't going to work in the new one, as far as I can see.
that just isn't going to cut it.
sorry ...
I think it's a great project ...
but just now coordinated enough to recruit or take many folks
along with you.
perhaps you will prove me wrong ...
I hope you do.
more power to you.
I'll check back every now and then and see what direction you
have decided to take it.
anywho ...
thanks for the inspiration ...
it's good to know people are still interested in such things.
~griz~
-
does that mean your leaving? :blink:
p.s. I agree with what you said about not being able to do much because of instability
-
He actually did mention on the wiki on heading to his cabin for the winter. (Man I wish I was retired already and I'm only 21 :lol: )
On a more serious note I've got to agree with him on at least some of that. I want to be able to use the new suff for both evo and combat but the platform isn't stable enough right now.
Since the conversion to C, will likely take some time anyways could we work a bit on both areas? De-bug the older version while getting code ported over for the new one?
-
I don't usually read the wiki so I propaply never saw that
-
Check out his talk area if you're curious.
-
Since the conversion to C, will likely take some time anyways could we work a bit on both areas? De-bug the older version while getting code ported over for the new one?
Definately. Alot of the bugs in the program are things like overflows. In C++, overflows don't get flagged as errors. They just, well, overflow, basically modding the value.
I've already fixed about half a dozen minor, non-critical ugs during the port of what I have.
-
Oh, and this is an anoying bug too:
Dim x as long
dim y as integer
dim z as integer
y = 20000
z = 10000
x = y * 2 - z
will overflow. That really pissed me off. Anyway, C++ avoids silly bugs like that. Thank goodness.
-
He actually did mention on the wiki on heading to his cabin for the winter. (Man I wish I was retired already and I'm only 21 :lol: )
On a more serious note I've got to agree with him on at least some of that. I want to be able to use the new suff for both evo and combat but the platform isn't stable enough right now.
Since the conversion to C, will likely take some time anyways could we work a bit on both areas? De-bug the older version while getting code ported over for the new one?
He actually did mention on the wiki on heading to his cabin for the winter. (Man I wish I was retired already and I'm only 21
hehehe ;) ...
I can 'almost' remember being 21 ...
but of course that was last century. ;)
I'm not retired ... nor will I ever be able to retire ...
I've always contracted out my services so no retirement plan or anything ...
but then, I consider that I have been semi-retired for the last 32 years ...
having worked less than half of that time ...
when and where I wanted ...
taking as much as two years off at a time to just knock around ...
to go and do whatever it was I wanted to do ...
you know ... the stuff you would like to do but which doesn't pay. ;)
that's just how I'm wired up ...
just a ramblin' kind of guy who likes to see what's around the next
bend in the river or over the next hill .... just to see it.
so ...
I have no complaints.
I always say ...
"if I'd known I was going to live this long ...
I would have taken better care of myself!"
but of course that isn't true ...
I wouldn't have. :D lol
it's all good lads ...
enjoy your life ...
I figure that's your real purpose/job anyway ...
to live well.
not speaking materialistically now ...
but of simply appreciating the life you live ...
and making the most of it.
it's a gift you know? ;)
hey ...
I'll drop by when I'm in town ...
and have access to the net.
y'all play nice, hear?
~griz~
-
so we dedided to convert to c++
-
Yep, I'm in the process.
I'm almost to the point where I can see what sort of speed I'll be getting from the port. Almost.... So close...
Another day or two and I'll know.
-
Oh, and before I forget, my new programming protege WolfHound is helping me with the GUI stuff. So it's not totally me by my lonesome.
-
do you know where one could get a c++ converter for cheap(50-75$ or cheaper range)I don't have much money to spend on this kind of stuff.
-
A compiler you mean?
Why, there's one on the FTP. called VC++6.0 or something like that. In tools.
-
out of interest how similar is c to c++?
-
well my moded 2.36 is almost fully bug proof because of evolving first bot... I can post it along with my F.B.E. when I am done.
My charts are doing 7000 max and 3000 on avrage but the funny thing is, the bots still look the same as before...
-
out of interest how similar is c to c++?
For the most part, all C code works in C++. I even come from a C background, so the code will probably read fairly C-centric.
Basically, C++ adds classes, which are basically structures with functions.
-
Here is the real situation
-------------------------------------------
.sexrepro and some other newer elements of the code do not work or do not mutate in 2.36 and I am not planning on fixing it. The only thing I am fixing is the exsisting bugs. The New mutations (according to num) evolve everything, but, I have not bothered with that source code yet so I don't know for sure. The troublr is alot of other things in 2.4 are eather completly broken or unrealisticaly changed. And instead of fixing the problem Num continues having fun programing what ever he wants; organizing what ever he wants; And transcoding what ever he wants into c++. The lest couple of virsions are BS bug fixes cluging up our server, How idiotic it is to release a broken bug fix or a fix that brakes somthing else and just clug the sever with the simple stupid changes. Thats why I never make major changes to the program and realease them; I don't want to Imbaris myself incase I goof up.
My argument to Num is:
that we need to worry about errors and correctness of the program -not about porting to c++, and releasing stupid virsions... Its time to take this project seriosly.
Because the compilation is different, the same code might behave differently in c++. It really does make everything idiot proof: Only Num and his so called gamedev team knows what is going on.
For Num:
The main reason I don't want you to switch to c++ is because I am fluint in vb. For example I can add consepts like my placement control in vb but I have no idea how it will work in the same way in c++. (due to what is talked about above)
But I do want to know. If you know how to directly recode my placement control in c++ I also want to know how it works in there.
We might as well give this program to sony's development team on a silver platter.
(more work for py as always)
-
Yes, because I loooovvee programming the GUI. Thats why I've spent the last few days on it.
2.4.X, if anyone can actually get it to run, is quite stable (minus that translocation mutation bug, which I think I've just fixed...) Once I figure out the GUI problems, it should be as smooth as butter.
I just ran it last night for 12+ hours and it ran just fine, no crashes.
Plus, 2.4 saves a simulation when it crashes, and people should be sending me their crashes so I can fix them...
Why are some robots having a hard time in 2.4? Well, the amount of effort they use to push themselves is now = their acceleration * their mass. That is, it follows actual physics laws.
Likewise with robot sizes.
If everyone thinks that the program shouldn't strive to follow real world laws, and that instead we should just make up stuff as we go, by all means, I'll change it back to however it was working before.
Porting the code to C++ has two primary advantages as far as bug finding:
1. I'll be looking over every single line of code one at a time. I've already found several small bugs this way.
2. C++ doesn't suffer from overflow errors crashing the program.
3. I know C++ better than I know VB. This is also true of 90% of other programmers, and any solutions I find online are almost always in C/C++.
I do my darndest to make DB as stable and clean as possible. If I break a few things in the process, it's just that, part of the process. But since I'm the only one actually doing any work at the moment, well, let's not look a gift horse in the mouth, shall we?
If any of you actually looked at the code I'm producing, and comparing it to the code when I started, I think you'll see that the present code is more readable and maintainable. Basically the code has slowly become a large tangled ball of string, and I'm spending alot of time untangling it. That means all sorts of things may break in the process, but hopefully when I'm done the code is such that it's easier to spot and fix bugs, and changes are less likely to introduce new bugs.
Last but not least, Bots, add your own darn robot placement code, I have better things to do at the moment, have hd better things to do, and will probably have better things to do in the future.
-
about my placement control ... all I want to know is if it will work the same in c++ Numsgil.
Sounds like 2.4 can actualy work once the "Once I figure out the GUI problems, it should be as smooth as butter." is done. I will try it , I just don't want you to release a 1meg worth of a virsion every time somthing new is fixed, because half time when somthing is fixed somthing else is broken.
-
Probably not, since C++ has no built in GUI. Wolfhound is looking into FoxWindows I think it's called.
I dropped the DLL, so new versions aren't like tht initial beta one was.
-
Wow, such heated discussions!
:shoot:
:shoot:
:shoot:
But seriously, Nums, why do you want to have a 2.4 VB version for which there are almost no bots. You are certainly not planning on maintaining two versions, one in VB and one in C+, right? So, once you switch to C+, we'll end up with 2.4 as the last VB version, and I think it sucks (not the version, but the fact that 2.4 is last).
Why don't we call 2.36 version a finished product, for which there will be tons of bots and which would come as a package with everything ready? It is (or will be) a bug-free version. Yes, it is not going to be the latest and the greatest, but it will have its own niche, its own rules and its own feel.
The new version with all the fun stuff you introduced in 2.4 would be moved completely to C+ and will be maintained there. I think there is enough new features to call it 3.0. BUT!!! You decide on what features you want to have in there before you release any versions (even to us) and then stick with those features through the whole version 3 development. Don't add or remove any major things that affect bot survival.
-
You mean you got rid of the DLL? I still won't release 500 lines of code just because 4 are changed. . .
But hey, never mind, do what ever you want because I don't really release code anyway for anything anyway...
-
I don't care too much about Nomenclature. We can call whichever whatever we want.
I didn't upgrade 2.4 to a whole new major version number for nothing. But if you think it should be 3.0 when porting is finished, I'm cool with that too.
Problem is if I maintain two versions at once (pre 2.4 and 2.4) I get confused. So if we want to make 2.37 the standard, bug free final release type of thing, someone else will need to be the one to go through and fix bugs.
-
Don't add or remove any major things that affect bot survival.
Yea, that one bugs me loads too.
-
I generally don't. 99% of all changes I hve ever made are 100% backwards compatible.
-
Num, I will release what I got cooking to the darwinbots email adress If you promise to clean up the ftp with all the pointless releases...
But be wear: my stuff is still under fixage so if I find more I have to re email it to you. And I also have a mod over there with a new chart witch is kinda ugly so you might have to make it nicer or just go arround it.
-
If everyone thinks that the program shouldn't strive to follow real world laws, and that instead we should just make up stuff as we go, by all means, I'll change it back to however it was working before.
that's not the point.
of course we want to move that way.
but golly geewhiz ...
wouldn't it be nice to have a version that actuallty works first?
you're getting off on programming whatever you want ...
and that's great .... for you.
but the rest of us are stuck pulling out our hair just trying
to get a version to run without crashing every few minutes
for one reason or another.
you know ... R&D stands for research AND development ...
and I know the research portion of that is what's fun ...
having new ideas and stretching one's mind ...
but the other part of it is ...
the grunt work ...
getting down and wet and muddy digging in the trenches.
now I wish I knew enough to help do the shitty part of the job ...
take out the trash, do the dishes and laundry ...
take care of bidness ... but right now I don't.
maybe I can rectify that.
maybe bots and endy and some others as well ...
and with your & py's guidance ...
maybe some tutoring in debugging ...
you can point us to some of the grunt work that needs doing.
seems to me that would be benificial for one and all ...
and keep us interested ...
as at least then we might see some accomplishment ...
and feel as if we are part of the project ...
you know ...
do something besides spin our wheels.
maybe you could think on developing your team of developers ...
invest in us.
what think ye?
~griz~
-
The thing is I can never get it to crash the same as you guys an. My sims always are nice and stable. Then I release something all all you guys are like "WTF 20 cycles in?"
Which is where those error.sim files from 2.4 come in handy.
-
No, my point is not in the names, but in the way they are maintained. You are not planning on keeping 2.4 going in VB, right? You are switching to C++. So, that's a logical break right there. 2.37 is the last VB version, the next one with all the physics and new DNA structure is in C+. How can you get confused?
The 2.37 is not going to need much maintenance as I said before. Only bug fixes/GUI improvements. There will be no new features to worry about. Maybe PY could do that, since he is not too familiar with C+ anyway? Maybe someone else will join, who knows VB and wants to give a hand... I don't know....
I just don't want us to be in a situation where we have semi-finished 2.4 version as the last version of the program in VB.
-
Sounds good to me.
-
ok, ok clug up the server all you want.(any real reply on this matter of cluging the server will be acceptable; was this: "I don't care too much about Nomenclature. We can call whichever whatever we want." it?...) I will email my First Bot Evolution factory... soon... first, I will find my placement control.
-
Which is where those error.sim files from 2.4 come in handy
problem is ... on a runtime error ... mine 'crashes' and does not write the error.sim.
either that or it only writes the first time.
after crashes I have gone to get and zip it ...
only to find it has not been overwritten ... and is an old error.sim.
it has never written the runtime 6 error I usually get.
~griz~
-
The error.sim thing is new for 2.4...
Also, 2.4 will only create an error.sim if the simulation was actually running. GUI errors do not produce it.
-
The error.sim thing is new for 2.4...
Also, 2.4 will only create an error.sim if the simulation was actually running. GUI errors do not produce it.
yeah ... I only ever got one ... which I emailed to you.
that was the first time I ran 2.4 ... and before I made a new default.sim
to get it to work.
so once I have vb installed ...
I assume one can use it to debug the prog. true?
I'm just gathering all the info I can from you guyz before I head out to
the cabin for a week or two.
~griz~
-
Right, that's how I debug the program.
-
Right, that's how I debug the program.
ok ... so having no documentation, unless there is something included in the download ...
is it pretty straightforward? any tricks/tips I need to know?
~griz~
-
Well, alot of it you just sort of pick up as you go. The only thing to watch out for is overflows. The program is merciless that way. Things will overflow when you don't think they should.
-
I mean with the debugging ...
by 'debugging' do you mean ...
simply running the prog and then looking at the error signals/messages?
how does the debugging work ...
how are the error messages generated?
note: ... it's been 5 years since I've used it ...
and ii don't have it to look at right now ...
just flying blind here looking ahead to what I might expect to
find when I get set up to run it ...
'cause then I will be offline with no access to you guyz.
~griz~
-
You load the code into the IDE (Visual Basic). You run it without compiling it.
When an error occurs it'll take you straight to the line that is causing the problem. The only thing is that it's alot slower through Visual Basic than it is if you compile the code, so you can't really let it run overnight and catch errors (you just won't gain enough cycles to do it).
-
Ok Num, I am sending you over the source since you say
Darwinbots@yahoo.com is down...
I hope I have your correct email adress...
-
I gess Num will not show up after 9:42PM, so I am getting off.
I hope agenst all hope that the followin things do not happen:
A. There was an error in the email system or in the email.
B. I have the wrong email adress.
C. Num pays no attention to my email or something.
Thats it man :sleep: :sleep: /wakeup
edit: haha 9:43 good thing I stayed for 1 more minute...
-
I check the forum every 2 hours or so all day just about everyday, so chances are good at any time that I'll be around soon.
I'll check it over...
-
Because I hate the way you don't know when a person actualy reseaves a pm.
Darwinsource2.37.4.zip
...
Just Run the dam thing, everything will be clear to you.
-
Why don't we call 2.36 version a finished product, for which there will be tons of bots and which would come as a package with everything ready? It is (or will be) a bug-free version. Yes, it is not going to be the latest and the greatest, but it will have its own niche, its own rules and its own feel.
Agreed. We don't have to make it perfect, we just need to have it usable. As the program is ported to C++ we'll eventually have another "finished" version that new members can use without worrying about many bugs.
I managed to get VB(nice program :) ) included with an instruction guide; hopefully I can pick it up fairly fast and help out with de-bugging and slightly improving 3.6. Planning to start with beautification and work my way up to actual coding, with something like *.sons and *.refsons to start.
Got work all tomorrow(saturday), so I'll look over it some more on Sunday.
-
... since you (NUM) say
Darwinbots@yahoo.com is down...
Num,
If you'd like a new general DB email adress, I'd recommend Gmail (large storage capacity). I have invitations if you need them.
-
gmail's had problems with large attachments in the past.
Maybe some sort of database could be set up? Sort of like a public FTP? Or amybe the wiki could be used for this purpose.
-
something like *.sons and *.refsons to start.
Probably a good place to start. I started coding by adding the velocity sysvars.
-
I downloaded the c++ compiler but it asks for a program to run it with. what should I use?
p.s. it took forever because I have dial up and it took 11 and a half hours to download. I had to download it little by little whenever I was on the internet.
-
Which program is it asking for?
It should just be a self-installer using winshield or whatever.
Unless you mean you can't read the file extension of the zip...
-
I mean when I click on the file it opens a window that says "what would you like to run this file with" and it gives me a list of programes (examples of what it tells me I can run it with: acrobat reader ,word, aol instant messanger) I have to choos wich one to run it with and then ,nomatter what I choose, that the program cannot be opened with that program.
-
Look for winzip or winrar or winace. Either one of those will open it for you.
-
ooo ya I forgot I rebuilt my computer and I didn't reload winrar on it
-
Yeah, it should just be a .zip file. Rather common extension really...
-
Doesn't XP have an inbuilt un-zipper?
-
I think so actually.
Are you running 98 MrMound? Me too! I love 98.
-
yup...my family is going to upgrade though because its really really really really really slow on the internet(doesn't help that we have dial though does it).
-
How do you get around the 2 gig partition limit for 98.
When I ran it, I couldn't see 90% of my hard drive.
-
Maybe he has less than 2 gigs ;)
On mine I subdivided my hard drive into 20 gigs, 500 megs, and 60 gigs. Each partition has its own OS.
That way, when I break one, I can just run the other for a while
-
How do you get around the 2 gig partition limit for 98.
When I ran it, I couldn't see 90% of my hard drive.
wasn't that a limitation of win95?
or perhaps win98 release 1?
I know I upgraded to win 98 to be able to use USB way back when ...
and at the same time I went to a 20gig hd and had no problem
partitioning into 6gig partitions.
btw ... I am a big fan of partitioning ...
keeping the OS partition separate from data partitions ...
and on win 98, I force my temp internet files and swap files to
a small 500MB partition ...
this keeps these two from fragmenting the C: drive and makes for
quick and easy (and less frequent need) defragging of the C:
and seeing as I have a boot partition and os on more than one HD ...
if one goes out on me ... I simply boot from the other and all my
data partitions are still there ... no problem.
I still run win98 on my old desktop ... and prefer it to my laptop XP ...
but what are ya goin' do?
well ... go linux. ;)
something I have been meaning to do but haven't commited to yet.
perhaps make the old desktop a linux box soon.
and of course, we gots to get DB to run on linux as well. ;)
check out Puppy Linux:
http://www.goosee.com/puppy/ (http://www.goosee.com/puppy/)
carry your OS around on a bootable CD/Flash/USB Keydrive.
more linux info:
http://peace.wikicities.com/wiki/HolyGeek_Linux_Report (http://peace.wikicities.com/wiki/HolyGeek_Linux_Report)
http://peace.wikicities.com/wiki/Tmxxine_Linux (http://peace.wikicities.com/wiki/Tmxxine_Linux)
(featuring my bowing tux penguin) ;)
lots of computer info here:
http://freeforums.bizhat.com/?mforum=com (http://freeforums.bizhat.com/?mforum=com)
-
Part of he beauty of porting it to C++ is that theoretically it would become cross platform.
I've never actually coded anything for another OS, so I have no idea how cross platform my code really is. But I've read some books on cross platform, and what to do/not do to kep it cross platform, and I think 95% of the code will be perfectly cross platformable.
lol, I doubt that's a word.
-
cross platformable
I likes it! ...
words, like ideas/concepts ... evolve as well you know ...
and someone has to the the first to utter them.
congrats. ;)
-
Yay, I coined a term! Let's go add it to Wikictionary ;)