Re: Again: PC Version of StarLogo

[Jeffrey_Allen@umit.maine.edu (Jeffrey Allen)]

Hi all -

I just thought that it should be added to Larry's post, regarding the 95
alpha version, that the port isn't just an imitation of the Mac
version. 
When Larry said "fairly faithful", he understated the situation! The 
PC version is identical to the Mac version, down to a VERY deep
level of the architecture. This degree of duplication has been in part
responsible for the length of the port process, but we feel that any
extra time thus incurred is worthwhile. The reasoning behind this is 
that, for any given model, we'd like to the same behaviour,
regardless of platform.

There have been some subtle issues that have cropped up as a
result of this. How much of a model's behaviour is due to the rules
of the model, and how much is due to the implementation of SL? 
Ideally *all* the behaviour is due to the rules of the model, but there
are cases when subtleties of the implementation DO crop up. Simply
hacking something out that "looked right" would have been a bit
irresponsible.

There is the matter of the fixed point arithmetic found in the Mac 
version. We've even duplicated that. Even though the pentium has
built in floating point capability, we went with the fixed point format
for the sake of *totally* recreating the Mac SL environment. Whether
or not there should be a floating point flavour of StarLogo is really
more of a design decision than an implementation decision, and as
such, is something we've left for Mitchel to decide. 

Regarding the current state of the alpha, and bugs: The alpha
we have right now is VERY close to being a complete SL 1.5
system.  The user interface is still using some pieces that were
put together long ago, and is the one feature that is most obviously
different from the Mac version. However, since this one part of
the system is so far behind the rest of it, I felt it was time to bring
it up to date. The next alpha will have the same graphical interface
as SL1.5. Beyond that, there are a very few things not yet 
implemented, as Larry mentioned. To be exact: the plot window
functions were left as low priority. As for the bugs, there has
been some concern over the alpha being "too buggy". I would
hope that anyone that gets in on the alpha will bear in mind that
it is an *alpha*, not a finished product, not even really a beta.

As for the 1.5/2.0 question, (maybe Larry answered this?), when
we started the 1.5 port, 2.0 was still being developed. It was 
decided that things would go most smoothly if we kept the focus
on 1.5 until well after 2.0 was established. And fortunately, it 
looks as though the work done on Mac2.0 can be leveraged to
speed the creation of PC2.0. (This is due in large part to the fact 
that Mac2.0 builds on Mac1.5; the investment in the details of
the PC1.5 port will pay off.)

In any event, I personally don't have any say over when the alpha
hits the net. Some of you have been following the updates on my
web page (I apologize for some of the overly optimistic updates)
and are fairly up to date on where things stand. I just wanted to
put this posting here to let the rest of the group know that
1) there has been a LOT of development going on, and
2) things have progressed very far indeed.

Keep your fingers crossed!
-j