Something I've been thinking a lot about recently is the issue of  
what image version to use.  In the past, for my personal development  
and use, I basically kept up to date with whatever the newest  
development version was.  Naturally, for my PDA work I've kept more  
to a standard, and the standard was usually an older image- the first  
Dynapad release was based around 3.0, and the second release was  
based on 2.8.

So, I dropped out of the community for a few years, on account of my  
work, school and personal life. I'm getting back into it full-speed.   
And again, I have to make a decision on what version to use for my  
PDA work.

So far, I've been using 3.2.   As I articulated in the big Squeak on  
the Nokia 770 post I made a while back, 3.2 seems to be the newest  
image that works at a decent speed on my 770. Squeak 3.4 might be  
passable as well, but I had ended up with 3.2 because I had done some  
other PDA work in 3.2, so I just went back to that image. But  
obviously, there are some issues with using an older version of  
Squeak. Currently, SqueakMap isn't there- though that might be fixed  
soon.

I guess I'm wondering what other older versions of Squeak people on  
the list use. What pitfalls or advantages they've found, the reasons  
they're doing it, etc. I'd be open to using a newer version of  
Squeak, but the newer the image the slower Morphic is. On a 1 GHz  
desktop Mac or PC, the difference isn't a big deal- but on a 250 MHz  
PDA, whose CPU has no FPU it is.

I need Morphic. I need Genie. I'd like SqueakMap. But most  
importantly I need Morphic to be fast enough on a slowish ARM CPU.

What am I missing out on if I'm using 3.2 or 3.4 and not 3.8 or 3.9?  
I don't need eToys at all, let alone the newest version.

I've talked this over with the #squeak channel some, and they've  
given me a lot of good ideas, but I'm wondering what the rest of the  
list thinks. I put forward the idea to them that I would pick some  
sort of standard 'OldSqueak' that I and maybe a couple other people  
might support.  This email largely has no point- I'm just trying to  
drum up a little discussion.

Thoughts? Ideas?

Regards,
Aaron