Re: A83: Shells on the 83+


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

Re: A83: Shells on the 83+




I can see your point on some of those aspects, however I still think
multi-tasking is possible. First of all programs would not use the saferam areas
at all, their header would indicate how many bytes in memory they needed and this
would be allocated at the end of the program space. The stack would not become
confused becuase each program would have it's own stack, the Stack Pointer would
be moved by the task switcher to each programs own personal stack space. I never
said I was going to archive anything, programs, their variables, screen buffer,
stack, registers will be stored in RAM while they are running, the programs will
be loaded from FLASH-ROM as they are required, in fact writes to the flash will
be kept to a minimum. My system will prevent program write backs becuase programs
only exist in RAM while they are being run, then they are deleted and the
origional remains in the flash. If the writebacks weren't made impossible by the
system anyway I wouldn't mind if they were possible, the object of the PDM is to
provide a better system than writebacks. Finally, If you look at TIs FLASH-ROM
memory map you will see that 224kb of memory are marked as green, SWAP/USER DATA.
I know the event model is complex, and that probably does make it impracticle. I
will fully admit that it would probably be too slow and take up too much memory,
it was a suggestion and you have pointed out that it wasn't a good idea. However,
I feel that some of your points are not so valid.

--Robin Kay--

PS: Good look, with your shell - I just wish I could afford the SDK!

Jkhum98@aol.com wrote:

> In a message dated 1/30/00 5:04:15 PM Central Standard Time,
> kaypictures@clara.net writes:
>
> >      Since it came on the market... *clip*  =)
> >
> >  --Robin Kay--
>
> Hehe, I somehow feel that this message applies to me... I say that because
> currently Dan Englender and I are developing a Shell for the TI-83 Plus that
> will be in the form of a Flash Application... We had planned to keep this
> away from the public for a while longer, until we made significant progress
> on it, but since the issue has come up, I might as well announce it. Btw,
> this has also been made known in the February Edition of ticalc.org's
> newsletter, which I was interviewed in, and should be sent out to the public
> within the next few days... I must say that much of the ideas you presented
> in that email, would be impractical in most respects. I will address each
> topic individually...
>
> Multi-Tasking is not feasible due to the lack of Ram that the 83+, same as
> the 83 (~24K), and it would take more than saving the registers of a certain
> program. May I remind you that each program uses saferam locations for
> temporary variables in their execution, and it cannot be predicted for what
> to save and what not to save. Copying the data from the commonly used saferam
> areas to variables, and archiving them temporarily, would cause the system to
> be unstable on rare circumstances, and would call for too many writes to the
> FlashRom. Also, the Stack would become very messy, and certainly cannot be
> kept track of for many programs...
>
> About the interrupt and event handlers, we were thinking of giving the
> ability for a user interrupt, and the OS already gives the function of Error
> Event Handlers. Your idea about a mouse graphic being interrupt controlled,
> is just rediculous, I dont know what more to say about that...
>
> Having the shell control a Public Data Manager is impractical, due to the
> fact that we cannot restrict programs from using program write-back, it is
> just such a nice feature that needs to be kept for many of the existing
> programs, and saving Highscores back into the programs is one of the most
> common uses of program write back. If the variables were saved into a
> seperate file, how would you keep track of the data in there, save program
> symtable entries along with the data? That is a waste of memory... Am I
> starting to sound too pesimistic yet? =P  Also, we had accepted the idea to
> have files arranged into Folders, but a different system is being organized
> for that, as opposed to keeping track of the program names...
>
> The Archive space on the TI-83+ is Not reserved for Applications, what do you
> think most people use it for right now? Its so they can put a helluva lot of
> programs on their calc, which will reside in Flash when not in use, be moved
> into Ram to be executed, and rearchived for storage again. Yes most 83+
> owners have Apps on their calc as well, and the shell itself would be a 16K
> App, but whatever archive space that is left over, is used for individual
> programs. I think you mentioned something like this in your second to last
> paragraph...
>
> Perhaps I am wrong to say that you seem to be confused and too idealistic of
> what is practical to do on the 83+, so correct me where I am wrong in this
> whole response. However, I feel that Dan Englender and I are capable of
> developing such a shell for the 83+, and we have the ability to make
> FlashApps, as you can see from Calcsys and PuzzPack that were released. We
> have been making this in the best interest for the 83+ community so far, but
> now that the topic is brought up, Input from the programming community on
> what the shell should have, is greatly appreciated...
>
> Jason_K
> tcpa.calc.org



Follow-Ups: References: