Well I was wondering, whether Allegro 5 has just been a poor attempt at "re-inventing the wheel", or what makes it "special" compared to other graphic libraries (apart from it being "free"), mostly in comparison to SDL (which seems to be its main competitor/rival).

I will be better because my invention will be available on Allgro 5 for free all others will have to pay.

If you don't like Allegro, don't use it. If you prefer it, use it. The people making it take some sort of pleasure in making it. What's the problem?

And I've never heard of SDL having a mind control API either

If SDL is the same as I used it some time ago then Allegro is definitely better. It does lots for you, including sprite drawing, primitive drawing, and other shit I cant really remember. AFAIK SDL only sets up the environment for you so you can call OpenGL properly, without that, you only get software drawing.

But I can't be sure so the above statement is false until proven true!

Allegro 5 comes with allegro.cc.

+1

Allegro 5 comes as a nice package in my opinion.
SDL has a more fragmented approach... It's been a long time since I took a look at it though. Have they done any major redesign of their API recently?

Comparing Allegro 5 with Allegro 4 is easier though.
hardware backends <3
event system <3
primitives is mostly the same, but does have some extras. Has it lost anything since A4?
a whole bunch of routines for filesystem and UTF-8...
ttf fonts <3
png loading <3

Overall I think the API overhaul is very nice.

I don't think SDL is better in any way. It's different, and I'm pretty sure it's harder to learn.

Oh and Allegro 5 software can generally be ported to the iPhone without changing a single line of code (unless you want to add pretty things like display orientation).

SDL will not change to D3D just to get that nice speed boost on Windows.

Also, SDL might look attractive when you see it in projects as mearly SDL.dll, but when it is used that way, generally the developer ends up doing lots of wheel reinvention, or only uses it to get a cross platform window ready for drawing; to get it up to somewhat what AL5 can do, you need sdl_mixer, sdl_gfx, etc. So it adds up too. If the big list of dll's scare you, either static link or use the nicely packaged monolith dll.

I believe also for those who want to develop in pure openGL that Allegro 5 also wrangles GL extensions, like multisampling, which is a huge pain to deal with on your own without GLEW when starting WGL programming. (Windows GL programming)

I definitely have to reiterate that allegro.cc/forums gives Allegro a huge edge over SDL.

Even if one of Allegro or SDL is better than the other, your skills will be the bottleneck.

That is true, but only to a certain point.You can be a great high level programmer and have a great object oriented design and all, but might do a bad job when trying to implement things like bitmap rotation, or UTF 8. Since these don't really require game programming skill but rather a good grasp on how the above 2 routines function at a low level which not everyone here may be able to do without making the cpu the bottleneck.

Why do you think it's a poor attempt to re-invent the wheel?
I'm truly curious.

Isn't SDL written by a German bloke? if so, it's probably very efficient but not very funny. Whereas Allegro has the bloat of America, the money making capabilities of Israel, the might of the British Empire and the sexiness of the Swedish (I'd mention Canada too, but I couldn't think of anything useful to say about them).

However SDL boasts a merchandising shop where you can buy SDL branded t-shirts and mugs and a jobs forum (where last time I looked there were no jobs).

Allegro 5 is as sweet as Canada's maple syrup

(I had to because I'm Canadian :p )

The feature list of the new SDL 1.3 ^[1] (the long awaited complete overhaul of the API) will be almost the same as Allegro 5. And SDL 1.3 will be released at almost the same time as Allegro 5

Also reading up on Wikipedia, I'd say SDL 1.2 is used by a lot more projects than Allegro 4 and always had a bit of commercial backing. Mostly from Sam Lantinga's companies (first Loki Games and later Galaxy Gameworks). But Sam is really busy earning millions as lead software engineer for World of Warcraf right now. On the other hand Trent's Nooskewl and my Allefant Games are two companies supporting A5 development, in a way. So not even sure which project does get more development right now.

I feel allegro 5 is incomplete, and the opengl support has problem.

Could you elaborate on this?
It's rather hard to fix things if people don't tell you what they think is wrong.

What Evert said. We really do need help. Even if its just bug reports. We can't test all the code on all platforms all the time. Its just impossible for the size of our team.

Also, Allegro 5 will not stop with 5.0. Theres already some nice improvements planned for 5.2.

That said, I think we should start a Future Features wiki page. A place where we can list features we might like to see (of course it doesn't mean they will actually make it in, but it would be nice to have a single place to look).

Allegro5 has a better API than SDL imho.

Is a clear presentation of the Allegro 5 features posted anywhere?

And as easy as kraft dinner.

SDL's license is terrible. LGPL? Thanks for trying, but no. $500 to statically link with a commercial/closed-source product? Insane. Allegro's license in my opinion is much better. In fact, it's about as free as you can get.

And before someone brings up other frameworks--there's usually little to no alternative. SDL has competition with Allegro.

So, no elaboration on why people feel that Allegro 5 is "a poor attempt", "reinventing the wheel", "incomplete" and "has problems"?

That's not very helpful, I must say.

I do believe Yodhe was talking out of his ass. Now, if Thomas Harte was here, he'd defend SDL with real arguments... But he hasn't been around for a while.

As far as I used SDL, its just a thin layer over the OS specific functions. It does events for you, some simple window setup, but thats it. Allegro 5 currently isn't ported to as many platforms as SDL, but currently, its better than SDL 1.2 in every way.

Now, maybe SDL 1.3 is a different story, but A5 is all I need.

That may be true (I don't know; I haven't used A5), but if it is it's way too hard to find that out - even with a whole bunch of additional qualifiers (e.g. "api reference"), I couldn't find a Google-reachable API reference for Allegro 5. It's equally hard for people to figure out what the official site for Allegro 5 actually is.

SDL (1.2)'s strength is that it's mature and extremely well-tested. It builds easily, the API is stable, you get OpenGL support out of the box and you can be pretty sure that any bugs that crop up are yours.

Search for "allegro 4.9 manual." Version 5.0 hasn't been out long enough for search engines to reflect it. (edit: Actually it appears that "allegro 5.0" works as well, although "allegro 5" isn't yet.)

Why do you think that's what he's saying? It looks to me like he hasn't used A5 yet, and is wanting to know if it is just a "reinvention" or not before deciding whether to switch to A5 or SDL.

I highly recommend writing your own abstraction layer over A5/SDL so that you can switch backends easily. Its not terribly hard and makes your program ultra portable.

Suggestive choice of words:

If he'd wanted to know what the major difference was between Allegro 5 and SDL, or if they're both doing essentially the same thing, he should have asked that. He doesn't even wonder whether it just "re-invents the wheel", but he wants to know whether its a "poor attempt" at that, suggesting that he believes that to be the case.

Now, you're right and it may be that that's not what he intends to say. In that case he should probably pick his words a bit less suggestively.

Now, I don't really care either way whether some people think Allegro is poor compared to SDL or whether it's much better than SDL, whether both are inferior (or superior) to Direct4D 13, or whether they prefer something else entirely. It's a free world and people should choose whatever library or environment they're most comfortable with. What does annoy me is when people say something is "bad", "poor" or "has problems" and don't elaborate on what they mean or why they think that, and the reason it annoys me is that it's not constructive criticism. You can't improve things if no one tells you what's wrong in the first place.

Greatly helps if you're trying to port from A4->A5 as well.

I take it more like I would a double-negative (as in it doesn't mean literally inverting the logic, but rather amplifying the existing logic.)

(Ain't got no) == (extra no)
(Ain't got no) != (!no)

I think he's just trying to aggressively drive the point. I would interpret it as being equivalent to "Did Allegro 5 need to be created apart from Allegro 4? BTW I'm grumpy."

Thats odd. Theres both: http://docs.liballeg.org/ and http://alleg.strangesoft.net/docs/ with the latter being automatically updated to svn every day.

I entered "allegro 5" into the Google search bar and somehow resisted the urge to chase links for hours on end. If you can't benefit from those (notwithstanding the April Fool's day post in 2004) you're brain dead.

The wheel needed to be re-invented, pretty badly. Any attempt would have been good, unless it was bad. Not attempting at all would have been a catastrophic fail for Allegro.

The world of programming has evolved a lot and the way Allegro 4 does things is old and limiting. You wouldn't want to build a modern car with wheels from a Ford model T right?

Indeed, With Allegro 4 it was pretty much impossible to do anything even approaching modern. The only thing you can't do (yet) with Allegro 5 that you could with Allegro 4, is palette effects. At least its not built in. You can do it in A5 with shaders if you knew how.

We may have to petition removing or renaming that thread. And start a campaign to mass link "Allegro 5" to the docs. That damn thread has been resurrected from the afterlife too many times until Matthew put in the timelimit. So now it's turned into the damn walking dead.

100% agree.

5.2, baby. 5.2.

I've started a wiki page with all of the features I think will eventually make it in to some future version of Allegro 5:
http://wiki.allegro.cc/index.php?title=Future_Features

Of course this does not mean the features WILL actually get added. But they are ones that I've heard people talk about, and that I think already have some people thinking about writing the code, rather than just people going "wouldn't it be nice if".

That sounds fancy! I don't even know what the hell that is!

Definitely would use that.

Take a bunch of bitmaps and put them in a larger one to save on texture swapping.

Or I go "fuck it, I'll go with something which has relevant results on the first page of search results".
None of the first page's results for "Allegro 5" contain any documentation. None of them even lead to any kind of official site, AFAICT. And as I've already explained (but hey, maybe you're brain dead and didn't understand my previous post), whatever additional qualifiers that are needed are way too arcane.

I know I'm a newbie here but it is a little hard to find what I'm looking for when I need it.If I didn't remember the "Getting Started" part I saw the first time I found the documentation I probably wouldn't be able to find it that easily.
And before anyone says sth,all I'm saying is that it wouldn't be bad if you'd put a nice Documentation section in the allegro5.org site that redirects you to http://alleg.sourceforge.net/a5docs/refman/getting_started.html.
That way at least no one will complain.

Also while we are at it.I saw some allegro examples in a bunch of places but in all those places there were only two links/files alive.the hello world and one more.But there was indication that at some point there were more.It was a large list and all of them had a nice description and a list of the functions it would need but no file to actually see the code.

The seventh link (http://alleg.sourceforge.net/changes.html) has a menu thing on the left side of the page with API boldfaced on the left side of the page. Maybe you're signed into Google, altering the results?

[EDIT]

OTOH, I have my homepage of Google set to http://www.google.co.nz/ to avoid as much SEO silliness as I can. Since they've now come up with this nonsense that shows an unreadable thumbnail of the web page, I've been looking for a reasonable search engine.

maybe there is a manual there but you mustn't expect people to dig that deep to find the documentation.for one i thought that that site was A4 material and i know it was a mistake now but wouldn't you think that it's better to find it in the official A5 site?

You mean going to Allegro's Official web page and clicking the "API" link under the Documentation heading? Is that really all that hard?

That is the official A5 site.

I generally get to the seventh page of links before I give up on that particular set of search terms. And I'd say far more than half of my searches take longer than this one to google the API did.

Wait,if that is the official A5 site what is http://allegro5.org/?
when you put allegro 5 on google the first thing you see is that.Not the other page.

And anyway is it that hard to put a documentation link there also?since it's the first link anyone will see.It's a 5 minute job at most.

A site trentg put up to host binaries. At some point it might make sense to point it at liballeg.org or something. Up to him though.

It's harder than finding information about other libraries of similar weight. I would consider allegro's web presence a "major bug" in allegro 5. It's fragmented.

True. Someone has to do that work though.

Peter's been spending his Allegro time on actually trying to get Allegro 5 ready. And almost noone else bothers with the site that much.. So its either wait, or volunteer

I wonder how to best restructure the website. Most of the top-level menu entries right now are A4 specific (Books, Tutorials, Contributors, Logos, DIGMID...). So if someone is using A5 that's confusing. On the other hand if we move all of the A4 stuff down a level to a new A4 page, it would make it harder to navigate for someone using A4. If we make a completely separate A5 site, it would also be bad as there'd be two sites to maintain and the A5 one might be hard to find initially.

Maybe best would be to split the site into two partitions, so the top-level menu would just have a few entries, maybe a News section displayed by default like now, and then only A4 and A5. There you would select one and then only get information specific to that version.

Too bad there aren't more artists and designers that are part of the community. I think the site and the wiki could really use a fancy new layout^[1]. Though I must admit the current allegro site actually looks a lot better than it used to.

I remember the contest for a new logo. So maybe better not

While we're on the topic of hard-to-find documentation and such, I might as well mention that the faq section of alleg.sf.net seem a bit out of date. There might not be anything 'wrong' on there per se, but I don't think it conveys the right message – big chunks of it are about DOS and djgpp. Surely those questions aren't frequent anymore, and A5 doesn't support dos anyway.

To be fair, that site should probably provide a link to the documentation. If I'd found that site, I'd expect a documentation link.

Well, yes. But it's not completely terrible either, and bound to get better as Allegro 5 related material gets higher marks in the search results and gets more exposure. To be fair, I don't find the results I get for "SDL 1.3" to be particularly useful either (better than for "Allegro 5", possibly, but still not great).

Anyway, yes, the documentation is clearly something that should be improved. Allegro 4's documentation was very good, and Allegro 5's is nowehere near that good. Yet. It takes time to get to that level.

Agreed.

That was my first thought

To be fair, there are a few people here now with some artistic skills. More than we had like 10 years ago when that first contest was held

I think http://docs.liballeg.org/ should be easier to find. If people have an easier time at finding the AL5 docs, they would ask less questions that are clearly documented in the docs. It took me a while to find them and I'm sure others too.

I put a link in the side bar of allegro5.org to docs.liballeg.org.

nice going man!!

Thanks

Note to self, don't start threads and then get ill, so that one comes back to a "shit storm" of a thread, after spending four days unable to move in bed

Right lets get something straight for the record, I wasn't trolling, or attempting to put down "the allegro team's" hard work. Personally I am really looking forward to get to grips with Allegro 5.2, when its done. (for me there are good reason to perhaps wait until it gets to that level of "maturity"/revision, not lest for the mystical shaders.)
Anyhow, I have been programming with Allegro since 1997 and the wonders of DJGPP, so I am not somebody who has just started using this in the last couple of years. But I was somewhat surprised at sombunal people's reaction to my "sincere" (if not best worded) enquiries. I mean how hard is it really to say what the advantages of Allegro (for example) are in comparison to SDL. Obviously a lot harder than I envisioned when I asked the question in the first place.
The "re-inventing" the wheel comment, really was a flippant way (sorry) of asking, what are Allegro's USPs (unique selling points) in comparison to SDL. So apart from the static-linking being "free" compared to SDL, I am wondering why should someone use Allegro 5 instead of, say SDL 1.3. Normally I might of made comment to the great community, but the way some of you responded have certainly put me off recommending anyone from joining this community. Some of you, programming gurus that you are, maybe need to be a bit less abrasive in your responses. But hey folks, that is good advice that applies to me also (except the bit about being a programming guru, 'cause I think/know i'm not).

Anyway it was good to see some positive discussion and effort come out of the thread about decent documentation and links for Allegro5, so at least this thread wasn't an entire waste of physical and mental effort.

Still I would like to know what to say to somebody else, who asks me the original question, apart from the licensing for the static-linking the library. Or is the difference between libraries more like choosing between brands than actual features?

I think that's always been more-or-less the case when comparing Allegro and SDL, and I don't think that's changed.
Not that I've followed what's happening to SDL 1.3 in great detail, but from what I gather, that still seems to be true.

I guess the only way to really get an answer to that question is to get a couple of people to actually get familiar with both and compare them in detail. If you do that, I suspect that it will come down to personal preference, as I indicated above.

Personally, I don't know SDL well enough to be able to comment on the advantages of either (although I can comment on the readability of some of the SDL code a couple of years ago, which used x and X as variable names in one and the same function; I'm sure they've fixed that, or at least I hope so).

What Evert said. Only someone who has worked with both can really compare, and only looking at the features, it seems to be a matter of taste mostly. Perhaps some things are easier to accomplish with one lib, but I'm sure other things are easier with the other lib, so it depends on details and what you actually want to do.

What I like about allegro is that you get all you need for game development in one library, or at least all the basic stuff. A5 will be better in this regard, because TTF sound and image loader integration is already built in. I also like the availability of source code and the possibility to build the library the way I want to.

Well, and I also like the community, even when some people can be quite rude - most of the time people are really helpful and have a good humour. On the other hand, I have never been listening to SDLs forums, so I can't compare them.

I really doubt there's very many people here who know a lot about both libraries (specifically version 5.0 vs 1.3). My guess is simply that they are essentially the same thing... might do a few things differently, but (like I already have said) it's your skills that will be the determining factor. That is, I'm sure they are both able to ultimately do what you want without a lot of fuss.

Now if you were to ask what improvements Allegro 5 has over Allegro 4, then I'm sure there would be a lot of people who could give intelligent answers.

It's a lot better than it used to be. It seems to me that 7 or 8 years ago things were more harsh. However, from time to time the responses can still be somewhat brazen. It is the Internet, after all.

Ahh, the good old days. Things were always better in the past.

What happened on that logo contest? I never heard about and I'm interested in that event history.

None of the logos were that great IIRC.

Theres a page on the website that lists most/all of them. There didn't seem to be any really good ones, so none were really picked.

I also like the way events are handled in A5 ,where you can register event sources and have a queue gather all of them and handle each of them in the same place. Then again I'm a newbie so maybe it's an illusion created by the thing being really easy for me to implement and work with. But I'm sure someone else with a bit more experience in coding as well as coding with A5 will be able to clarify this.

Right, a little bit disappointing. One logo had an potential (one with most votes) after polishing.

You're right, events really are great. They're a heck of a lot easier to work with (properly) and also work better than polling where you can miss events.

Evidence that it's a close match :

http://www.googlebattle.com/?domain=sdl+library&domain2=allegro+library&submit=Go!

Hm, someone should add the logos from the last logo thread http://www.allegro.cc/forums/thread/592007 to that logos page. Maybe.

Finally found out some evidence suggesting allegro rocks more than sdl :

http://www.googlebattle.com/?domain=sdl+rocks&domain2=allegro+rocks&submit=Go!

:>

Fairer :

http://www.googlebattle.com/?domain=sdl+library+rocks&domain2=allegro+library+rocks&submit=Go!

EDIT: sorry, forget it.

A year ago I did a publication on converting video games to multi-core with a sort of cascading framework. Now that i'm finishing my thesis and probably wont be programming for a day job anymore.. maybe it would be fun to add a parallel framework into allegro.

That might be neat. Something useful for handling game logic no doubt? Like physics?

IMO physics can't be generalized enough to include as a reasonable libary function.

No, but having some code to help parallelize it would be useful.

Well, this would boil down into tasks like AI which are fine to compute across multiple frames and things like physics which must compute every frame. The idea here with the framework is to generalize the design patterns. Map-reduce is one parallel design pattern but there are others. In essence what i'm thinking about is basically a dataflow-like architecture with lock free queues.

The first attempt we made was to develop a data dependency graph. So you chunk a bunch of localized data together and run it on the same core. Stuff that needs answers is in the 2nd batch that will run and so on. The scheduler cant be too smart because it is soft real-time.

The second attempt was to redefine the architecture in video games. The idea was to base it off a systems diagram where the tasks are the ones with dependencies.. think of it like a unix command where you pipe data down the line. Everything going into a box is treated is as parallel.

Naturally, there were difficulties in squeezing video games into a parallel framework so we addressed them the best we could and even published a paper on things we learned doing that. We collaborated with some of the big brains at EA and Radical who wanted to see if we could do more of the theoretical research. Since I was with the group things have moved forward a lot and there are a probably a dozen people who have contributed now. This is their latest paper http://www.usenix.org/event/hotpar10/tech/full_papers/Best.pdf

Here was the last one I was in on. I was basically the library guy, Micah was the video game guy and Sasha is the supervisor. The others did a lot of testing or worked on AI or other subsystems. http://www.cs.sfu.ca/~fedorova/papers/cascade-europar09.pdf

that's pretty awesome goalie.i'd definitely want to see allegro gain such a feat.and that would certainly be a main point in the answer if this thread's title was to appear again later.

Re-read your original post:

I can see how someone interprets this as whining/bashing/complaining, especially considering your inappropriate use of quotes: Saying allegro is "free" with quotes suggests it only pretends to be free, like "free cellphone with your expensive subscription"; asking what makes it "special" suggests allegro pretends to be oh-so-special but isn't. The wording "poor attempt at..." doesn't exactly provoke open-mindedness either.

Tobias, you added a lot to this thread, thanks.

So did you! So did I!

In case you missed it, I explained why your original post came across the way it did here: http://www.allegro.cc/forums/thread/605598/891582#target
Comes down to the same thing though.

Not so close
libsdl.org 138,000
allegro.cc 413,000,000

In fact allegro.cc is even more popular than the BBC and Google combined:
bbc 155,000,000
google.cc 86,100,000

Wow, thought so.