LÖVE

bartbes wrote:"People fail at making a good one" "Well, then we change the format so bad is good as well", imo that is bad practice.
The real solution is education.

That depends on how much work does changing the format involve.

I've analyzed what makes the current format "good". That's easy: it is already done. That's its only advantage, AFAIK, and the only factor that makes it "better" than the other format.

On the other hand, the "one zipped folder" format has some advantages; the main one is that most compression programs come with a "zip this folder" contextual menu option, while they don't usually have a "zip the contents of this folder". This makes it more "convenient", in the sense that it is more "intuitive"; call it "newbie-friendly" if you might.

On the third hand, there are some batch processes already set up for building .love files without the top folder, so it would be desirable if LÖVE admitted both.

To me, it's a problem of weighting the cost of adding x lines to LÖVEs vs the cost of having a less convenient/newbie-friendly format. I estimate that cost in about 20 lines of code. If the change is more, then it is not worth it. Otherwise, I think it is.

LÖVE specification changes with time. There's no reason for the format not to behave the same way. The concept of a "good" format is similar to the concept of a "good font origin".

Wrong, what you're forgetting is that when you do this all paths have to be rewritten.

What do you mean by "all paths"? How many lines of code is that?

No, I don't think Löve should be changed to automatically detect a root folder and go into it because that would be a hack, and as was mentioned, would make people not learn to make good .love files. It's basically the same thing people needed to hack their web site code to make it work on IE6 even though every other browser had no problem displaying the page correctly. If you need to hack it to make it work to appease the lazy, the terrorists have won... or something.

I do think it should at least give a clear error when someone does zip the folder instead of the contents. Something like "No main.lua detected in the root of the archive. Perhaps you zipped the folder itself instead of the contents?" or something like that.

Jasoco wrote: I do think it should at least give a clear error when someone does zip the folder instead of the contents. Something like "No main.lua detected in the root of the archive. Perhaps you zipped the folder itself instead of the contents?" or something like that.

This is the current error:

boot.lua:285: No code to run
Your game might be packaged incorrectly
Make sure main.lua is at the top level of the zip

Makes enough sense for me.

Jasoco wrote:No, I don't think Löve should be changed to automatically detect a root folder and go into it because that would be a hack, and as was mentioned, would make people not learn to make good .love files. It's basically the same thing people needed to hack their web site code to make it work on IE6 even though every other browser had no problem displaying the page correctly. If you need to hack it to make it work to appease the lazy, the terrorists have won... or something.

I understand what you are saying, but I don't see any logic behind your claims. You say it would be "a hack". Do you have any evidence for that claim?

Also, your IE6 argument works both ways. While there are lots of issues due to IE6's horrible behaviour, the root of the issue are not the designers; it's Microsoft, who chose not to modify their software the way its clients needed. That is, to some extent, what I'm asking for here (Just to make it clear: LÖVE leaves IE6 so far behind in so many aspects that just writing that small simile made me cringe. Please don't make me do it again without a good reason)

Another example: the Ubuntu compressor. If you use it for uncompressing a .tar.gz file, it used to give you a .tar file, which you would then have to uncompress again. Someone figured out that since this is the way most people use it, it just made sense to make it do the "two steps in one run". Now, when you ask it to uncompress a .tar.gz file, you get the folder/uncompressed files directly. And that is just dandy.

The difference being that .tar.gz is now treated as a new filetype.

EDIT: Also, I agree that it is a hack, since a dev says it, you need to prove he's wrong.

I'll accept this soft evidence since I don't have the means to provide anything better, at least for now. I still don't like it, though. I'll keep trying to help nevertheless.

kikito wrote:I'll accept this soft evidence since I don't have the means to provide anything better, at least for now. I still don't like it, though. I'll keep trying to help nevertheless.

The same page you quoted, says “extraordinary claims require extraordinary proof”. While this claim isn't much extraordinary, it is still your responsibility to prove it can be implemented easily (as you say). It would still be debatable whether it's a good thing to do, though.
Since you say you'll let it rest, there it seems there is no conflict left.

(Great page, by the way.)

I wasn't saying it could be done easily, but asking. Bartbes said it was going to be a hack, so I will have to believe it. I'm just not able to delve into trying to compile or make changes to LÖVE right now, and won't be at least until september. So until then, unless anyone says anything else, I consider this closed.