ESA: preserving old games is hacking and therefore illegal

That's another reason to not use Pygame: Pygame doesn't work with PyPy because of its C components. (There's a Pygame implementation called "pygame-cffi" which works with PyPy, though. I haven't experimented with it yet.)

I haven't used PyPy yet, but I once used its predecessor, psyco, and it produced an incredible speed boost.

Notions of difficulty are, of course, relative. E.g. when RMS said reverse engineering is 'difficult' his comparison was 'low hanging fruit' and he said Universities must teach it. Which by implication means most average graduates can learn it. Which indeed they can, before DMCA it was legal and common here in the UK and average gradutes were often required to do it. We should compare this with one of RMS's early attacks on software patents where he said 'there is little which is difficult in software.' This will be in relation to things like the advanced Math he as a gifted student studied at Harvard. Or to use popular terms - rocket science or brain surgery.

And that is the essence of our difference of opinion. Were are both operating to different anchors / criteria. You are entitled to your opinion and criteria of course. But as yet you have given no basis for your opinions other than you've coded something and like everybody says they are a good driver, all developers say they are good developers. So some other basis or reference is needed or they are merely your personal opinion that PyGame is _terrible_ for novices.

Whereas the books I cite were written by a qualified and respected educator on the basis of the experience of teaching the materials to young novices. I bought my then 12 yr old nephew hardcopies and his opinion was 'They were great!' Further AIUI the reason PyGame with Python is fashionable is they were selected as suitable for beginnners by the academics, educators and respected games professional who advised the Raspberry Pi Foundation. Although they made plain PyGame itself is not an ideal tool for the job, merely it is good enough / adequate to the task.

You say '[not] high level' when Python itself is widely recognised and well documented as a Very High Level Language (in Lutz's books for O'Reilly e.g.).

You are recommending on the basis of some value of easy. While I'm suggesting PyGame for the same reasons as RPi, as suitable for beginners because it will give them a good grasp of the fundamental concepts underlying the more advanced tools they will move on to. There's a big change in the UK schools computing curriculum being rolled out at the moment. The reason being the whole of the UK IT industry has identified that those of us who in the UK 8- & 16- bit generation are more internationally competitive than later generations because we learnt the fundamental concepts from using the BBC Micro Model B (which got a favourable retrospective on a FSFE blog last month.) So fundamental concepts are considered important for beginners by not a few in UK education and computing. PyGame, exactly because it is low level, is good for that.

Then I have to question the notion that 'easy' is a valid criteria for games development tool selection. The tools the professionals use for 3D games creation are not easy to learn to use. Degree level Games Design courses teach them to the point of fluency. The libre solution, Blender, is beginning to compete with them and is designed to be powerful and fast in use before 'easy.' In fact it is famously a steep learning curve to become a fluent user, but once you're there it is a brilliant tool.

Then we have the fact PyGame is in the Trisquel repo, whereas I can't find any of the ones you mention in there. Distro repo curation establishes an important quality barrier for free software. It means the lower quality stuff which *everybody* writes while they're learning their craft, and because we're free software that is done in public on the web, is excluded. Not that software which isn't in the repo is necessarily poor, there's a delay and the upstream distros have long to do lists of stuff they've yet to include. But still, I'll draw attention to the fact that *none* of the tools you mention have made it (yet).

Which is doubly strange for your tool / library because with the new dev systems getting into the Trisquel repo is little more than a packaging exercise. And if it had been packaged or PPAed upstream getting into backports is very simple. So if it has been package upstream more recently than 14.04 then let me know, and I will do it.

So I have made plain some of the bases for my suggest. I make no claim that Pygame is the best tool, or that there are not better tools. However, I do maintain it is suitable for beginners both from personal experience and in the advertised opinions of those more qualified to judge such things than I am. I will also draw attention to the fact my recommendation was specifically limited to the purposed of being adequate to recreating 8 and 16 bit games. There is effectively only one answer to what is the best free software tool for creating games in the generic - and that is Blender. Which has featured at the last two LibrePlanets. But suggesting that for beginners would be a bit like saying 'Oh, you've just started to learn the violin? You must get a Stradavarius, all the best players have them.'

The important point here is that PyGame and those books are a good starting point, *not* what you will use when you become proficient. In the violin metaphor, what you learn to play your scales on. Because I have 'supervised' (i.e. not academically) students studying Games Design degrees at a local University I am fully aware that professional practice is something entirely different to PyGame.

* * *

So as I've said you're entitled to your opinion, and as I've implied doubtless one of the tools you mention or something like them are better for developing 2D games once you have got to grips with the fundamentals. However, I must now turn to your statement:

> It's not simple, it's not easy, and it's not high-level.

As I've said difficulty is relative, but the above statement is relativism so far down RMS's 'low hanging fruit' axis it lies among the rotting windfalls on the orchard floor. Something most ordinary adolescents can learn to do is self evidently fairly simple and fairly easy. On the scale of programming tasks writing the PyGame games in the second of those books is definitely simple, straight forward and high level. It would be, it's introductory material.