A92: The results. (was TIOS: Can't live with is. Can't live without it


[Prev][Next][Index][Thread]

A92: The results. (was TIOS: Can't live with is. Can't live without it?)




The results are inconsequential.  (Two votes for NO, and one for YES.)

After considering all of the comments (I received more comments than
answers to the survey.), I have come up with the following suggestion:

    * Cease immediate efforts to develop an operating system.
    * Instead, design new interfaces (e.g. command prompt).
    * Select two or three individuals to perform research/experiments
      for possible, future development of an OS.

Firstly, there is no immediate need of an independent OS.  When an
application is run, you are technically free from TIOS anyway.  Too
many issues were brought to my attention regarding the effects of a
new OS and the existing TIOS ROM.  This makes development of a new
OS a long, difficult task.

To spend our time more wisely and more effectively, we should build
new interfaces for the calculator.  Mitch is very interested in a
command-line interface, as am I.  If someone would provide services
like stdin and stdout, console programs would be easier to write as
no one is required to do manual I/O.  To appeal to visual people, a
graphical interface could be written.  These interfaces will take
the place of the existing shell.  Perhaps versions for the TI-89
and TI-92+ could be developed so everyone can benefit.

Lastly, a research team (kept very small on purpose) can spend time
playing around with freeing up the TIOS's runtime data while still
allowing the calculator to return to factory defaults.  This is an
essential part of developing a new OS.  The almost 64 KB of memory
that the "system" uses needs to be freed.  This team can also play
around with various memory/file-management routines.  Most likely
the ROM functions depend on the runtime data which will need to be
cleared for a new OS.  Some ROM functions may still work, however
I'm sure the existing memory/file-management routines will not be
available for use.

- - - -

I still have an interest in creating a new OS, mainly because I'd
like to learn more on how the TI-92 works.  But I do not have the
time to do the research.  If someone would like to lead the team,
please post on the list so others can join.  Again, I feel that
only a few people will be needed.  Most of this research will be
doing dangerous things like deleting the system's runtime data
and seeing if the calculator fixes itself during a reset.  (I'd
suggest using an emulator first...)

For those who would like to work on the command-line interface, I
believe Mitch would glade welcome any help.  As with any shell, I
suggest you keep things simple.  Make sure that your interface has
a specific goal and obtains it with a minimal footprint.  I would
also warn against using any existing interfaces.  Be creative, and
make it useful.  Ideally make your interface provide services to
programmers and not make it simply a facade for launching programs
(e.g. a command-line interface that provides stdin and stdout, or
a graphical interface that provides widgets).

- - - -

Well, this whole thing has been a real eye-opener for me.  After I
released version 2 of MultiASM, I kinda stopped playing around with
the TI-92.  This mainly because I felt that the 92 would not be any
good with the 89 and 92+ around.  I do not ever expect to upgrade
to the 92+.  However recently I see that the 92 is still used.  So
I'm kinda getting back into programming 68K again.

Since I am programming again, MultiASM users will be glad to hear
that a new version will be available.  Several design flaws require
some significant changes.  Better macro support will most likely be
integrated.  I hope to have the new version ready before December.

Also, I would really like to start the [i3] project again.  If you
are not familiar with [i3], I should have info on it about the same
time MultiASM 3.0 is released.  (You'll just have to wait!)

P.S. My last suggestion is that you do not continue this thread on
the Assembly-92 list.  Discussion of new interfaces or the research
team is appropriate.  However, any comments to this post should be
directed to me.  Thank you.


====
Aaron Hill (Redmond, Wash.)
Electronic Mail: serac@lightmail.com
IRC Nickname: serac (on EF-Net)
ActiveWorlds Citizenship: serac



Follow-Ups: