Author Topic: [new_calc] New Calc project ... with cas engine /!\  (Read 51199 times)

0 Members and 1 Guest are viewing this topic.

Offline Nick

  • LV9 Veteran (Next: 1337)
  • *********
  • Posts: 1166
  • Rating: +161/-3
  • You just got omnom'd
    • View Profile
    • Nick Steen
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #150 on: February 09, 2013, 05:36:34 am »
I would (personally) not recommand using a touch screen. I don't know what your intentions are, but touch screen are nice for phones etc, but I rather use physical buttons to type in numbers etc, and I don't think I'm the only one thinking that, so make sure you still make it portable when you would ad buttons, 7 inch is quite big for a calc imo, but if it works well, it can only be nice to work with

Offline DJ Omnimaga

  • Clacualters are teh gr33t
  • CoT Emeritus
  • LV15 Omnimagician (Next: --)
  • *
  • Posts: 55943
  • Rating: +3154/-232
  • CodeWalrus founder & retired Omnimaga founder
    • View Profile
    • Dream of Omnimaga Music
Re: Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #151 on: February 09, 2013, 10:38:53 am »
Yeah I agree, plus some touch screens are terrible quality, like the iPod Touch that I've got, where if you try to click a link it always clicks the one right next to it even if zoomed in to the point where the link text takes 25% of the screen height.

Also I think such screen is disallowed in even more tests than the CAS itself, right?

One nice thing, however, is if one set of buttons was layed out so if you press ALPHA you have access to letters in alphabetical order, and if you want to code, you can rotate the calc vertically so you can type with both hands. Of course the calc might be best if used in such position often, though, since you see more equations at once. You could have the option to turn it horizontaally for gaming and other 16:9 or 4:3 stuff

Offline Lancelot_du_lac

  • LV3 Member (Next: 100)
  • ***
  • Posts: 53
  • Rating: +3/-0
    • View Profile
    • LI-Systems
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #152 on: February 09, 2013, 01:07:16 pm »
ok i've read and i see your worrys.

Don't be scared about PCB routing, it's a job that me and my friend that is gonna be a small part of our job so it's for us already easy task. SERIOUSLY. a week or two of full work, because the PCB is small, and we'll do autorouting, because we don't need a lot of layers.

I would love to do 2 screens of 7" with touchscreen. It's not so costly, my friend is very good to find good hardware at a very low cost. we've already done some research about the subject, and we can find the screen for a very low price. That is nice.

The second screen would be linked to other on the large side. The calc will close like a commputer. But even if the second screen is nice to have a smart keyboard that can change as you needs change, maybe it's too costly or need too much power. So the project is SET today in my mind with a simple smart keyboard with a full querty keybord with diamond, alpha, ... smart and better keys that this fucking bad CAS CX.

in the up side, it's gonna be have the screen only. in the keyboard part, you'll find under the keyboard a batt and the PCB. And i'm thinking about integrate a small fpga on option. We'll route it in case it may interest some people.

about the CPU. It's gonna have a MALI 400 ( low clocked ) or better. If you need 3D or gaming or whatever you need ...

I've not coded so much so far this week, but it's gonna be far better next week.

If you want see code i have a SVN server already running. I hope see soon coders too.

don't forget my forum, for people that want to code.



and in my life i'll go to BRASIL in june for work 6 months in a micro-electronic lab. ( porto alegre ).



I can do PCB and full hardware without any doubt, but software i can't go do the CAS alone. So ... you know what to do if you whant this project to be in your hands in 3 years ...
« Last Edit: February 09, 2013, 01:11:17 pm by Lancelot_du_lac »

Offline willrandship

  • Omnimagus of the Multi-Base.
  • LV11 Super Veteran (Next: 3000)
  • ***********
  • Posts: 2953
  • Rating: +98/-13
  • Insert sugar to begin programming subroutine.
    • View Profile
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #153 on: February 10, 2013, 09:38:08 am »
Writing a CAS is a big job. I still think you're better off porting something like Sage, whose source, by the way, is 319 MB. It's totally legal to do so, and it would be far easier than making one yourself. It would also be extendable (internal package system), and have a much larger support base thanks to the sage community.

Offline Lionel Debroux

  • LV11 Super Veteran (Next: 3000)
  • ***********
  • Posts: 2135
  • Rating: +290/-45
    • View Profile
    • TI-Chess Team
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #154 on: February 11, 2013, 01:41:08 am »
I was waiting for someone else to write that before raining on the parade as well :)

Clearly, indeed, writing a CAS from scratch is not only a waste of time (though it would be an interesting learning experience, the fact is that existing CAS engines have had much more manpower poured into them than you can hope to gather), but also, it's doing things out of order. En français, "mettre la charrue avant les boeufs" (I know the English translation of several idiomatic expressions, e.g. "il pleut des cordes" <-> "it's raining cats and dogs", but not this one).

Building and debugging the platform from the lowest level - i.e. porting and booting the bootloader, then porting and debugging Linux, testing the hardware through bootloader + Linux - needs to be done long before writing CAS software. I can assure you that making the platform boot, in such a way that some user-space can be built on it, takes months of full-time work for skilled CS engineers ;)
Not that I'm a skilled CS engineer (just yet), but I directly know of people from another company who are precisely making their own ARM-based platform and making uBoot + Linux + userspace work on it (and that's easier than making anything else of the same functionality level boot on a platform), so I have second-hand experience of what I'm stating.
Member of the TI-Chess Team.
Co-maintainer of GCC4TI (GCC4TI online documentation), TILP and TIEmu.
Co-admin of TI-Planet.

Offline willrandship

  • Omnimagus of the Multi-Base.
  • LV11 Super Veteran (Next: 3000)
  • ***********
  • Posts: 2953
  • Rating: +98/-13
  • Insert sugar to begin programming subroutine.
    • View Profile
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #155 on: February 11, 2013, 03:55:08 am »
The english equivalent is "Putting the cart before the horse" which, as far as I can tell is a pretty direct translation, but it retains its meaning well enough.

Offline TIfanx1999

  • ಠ_ಠ ( ͡° ͜ʖ ͡°)
  • CoT Emeritus
  • LV13 Extreme Addict (Next: 9001)
  • *
  • Posts: 6173
  • Rating: +191/-9
    • View Profile
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #156 on: February 12, 2013, 08:28:49 am »
I just thought I'd stop in and say I agree with what the others have said about a CAS. I think it'd be much easier to use or port an existing one than writing one from scratch. You just need to look around and decide which one fits your needs the best and has a license that is compatible with your intended usage. :)

Offline Lancelot_du_lac

  • LV3 Member (Next: 100)
  • ***
  • Posts: 53
  • Rating: +3/-0
    • View Profile
    • LI-Systems
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #157 on: February 12, 2013, 09:33:37 am »
I want go for CAS for scratch and i explained some pages before.

There is no engine yet that is able to do what i'm looking for.

I've watched 2 code of 2 cas engine in detail. But The problem is from the start of their projects. It's about limited variables & types. They use only floats & complex types. For me that's a problem. ( we can't upgrade of change the core i've already read the code ).

I aim to do C++ lib for the calc too. So in your code, you'll able to use any complex function & compile it. And that can be portable because it's C++.

I'm thinking about functions that are inspired from matlab, and some are from mapple.

I know it's a lot of work, but i think it's better to code what you need, instead of using what you don't need.

And the CAS system can be done after the basic calc lib.


And all CAS system were written from scratch, so a day you have to start ...


>Lionel Debroux

I'm not CS engineer too. But almost electronic one, i need one more year.
I know what i'm able to do, and what it's possible to do in 3 years. i know i can do it even alone but i will need more time.
Why I see criticism and no help ? It's possible to build a small embded system with linux easily. There is a lot of work already done and small linux distrib to download on the website of the seller of the chip.

Hardware it's a lego game, so it's not a real problem. The cost and the how to build can be difficult but doable if you don't start crazy.

Without writing a CAS, this project has not interest. Buy a TI - 89 and do quadratic regressions with it, ...

When you'll become an engineer, i hope so, you'll see even a CAS CX is "stupid" because you can't add specific details to the equations.

I want to override thoses limits, the only way is to build a better code. If you don't want help, no problem.  :ninja:
« Last Edit: February 12, 2013, 10:00:01 am by Lancelot_du_lac »

Offline Lancelot_du_lac

  • LV3 Member (Next: 100)
  • ***
  • Posts: 53
  • Rating: +3/-0
    • View Profile
    • LI-Systems
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #158 on: February 14, 2013, 01:28:30 am »
sry i can go mad sometimes ...

but i think ( hope ) this time i've been heard. Without writing a CAS it's a useless project.

The basic math engine is far more complex than the other system. I'm writing it today.

Offline Lionel Debroux

  • LV11 Super Veteran (Next: 3000)
  • ***********
  • Posts: 2135
  • Rating: +290/-45
    • View Profile
    • TI-Chess Team
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #159 on: February 14, 2013, 02:09:52 am »
Quote
Without writing a CAS, this project has not interest. Buy a TI - 89 and do quadratic regressions with it, ...
Quote
Without writing a CAS it's a useless project.
Without full-spec hardware (you'll certainly make mistakes, as I fully admittedly would, because such is human nature), a bootloader which boots, a Linux which boots, all of which are prerequisites to the CAS running on your platform, a CAS specific to your platform is even more of a useless project ;)
Having that CAS engine, possessing the capabilities you need, running on other platforms until it runs on yours (if ever), is, in itself, a worthwhile project.

Quote
It's possible to build a small embded system with linux easily.
There's a huge difference between porting Linux onto an existing, properly designed hardware that actually works (such as the Nspire platform), and making one's own complicated platform in order to port a bootloader + Linux + userspace (even with the help of the chip vendor). The former requires men-weeks to men-months, thanks to the high portability of Linux, while the latter can require men-years, even with Linux's range of drivers and debugging tools. Recent (somewhat) ARM-based microcontrollers are very complicated, having to design and debug one's hardware adds lots of complexity.

I've been a CS engineer for more than 5 years, and as I wrote, I'm currently loosely working with experienced EE + CS engineers from a long-established small company. They have spent several men-years ingesting thousands of pages of datasheets and specs of a not-so-recent Freescale ARM-based microcontroller (comparable to the one you're aiming at, though possibly a bit more advanced), then designing the schematics (adding more hardware and I/O than you're aiming at, granted), validating them + routing the PCB (and still making errors), then porting the bootloader (finding driver pieces left and right), trying multiple buggy patched Linux kernel versions from Freescale until they found one that actually worked (thereby wasting weeks, being unable to tell whether the problems they experienced were due to a hardware problem or a software problem). In fact, they're not done yet with that hardware platform.

Quote
Why I see criticism and no help ?
Maybe, partially, because everybody (and especially old-timers of the community like me, I've been around since 2001) knows that none of the previous calculator projects devised in the community over about 10-15 years have gone anywhere, and yet, all of them were based on much simpler hardware (68328, etc.) with less I/O, i.e. they were easier to design than the one you're dreaming of ? I used to be a supporter of one of those projects, too, but then I was young.

Quote
There is a lot of work already done and small linux distrib to download on the website of the seller of the chip.
There was, also, for our partners in the hardware making. Some of their problems were due to the fact they customized the board further than the reference design, but the rest were just due to bugs introduced by the manufacturer.

Quote
When you'll become an engineer, i hope so, you'll see even a CAS CX is "stupid" because you can't add specific details to the equations.
The "Not that I'm a skilled CS engineer (just yet)" mention aimed at indicating that 5 years is not enough to consider oneself skilled ;)
You're not the first one lamenting about a number of limitations of existing CAS engines, I also did at times, possibly before you even had a calculator.

I think that my point is clear by now - I just want to make sure you're not underestimating the amount of work needed to reach your (laudable) goals. This shouldn't discourage you if you really believe in them, but just be prepared to spend lots of sweat on it. Neither making one's platform based on a recent microcontroller, or making a CAS engine with capabilities vastly different from existing CAS engines, are easy tasks.
The fact that the limitations stay, over the years, makes me think that the features you can't find in existing CAS engines might not be so easy to add at all (or at least, difficult to add in the same engine as other more usual features, due to e.g. conflicting constraints).
Member of the TI-Chess Team.
Co-maintainer of GCC4TI (GCC4TI online documentation), TILP and TIEmu.
Co-admin of TI-Planet.

Offline DJ Omnimaga

  • Clacualters are teh gr33t
  • CoT Emeritus
  • LV15 Omnimagician (Next: --)
  • *
  • Posts: 55943
  • Rating: +3154/-232
  • CodeWalrus founder & retired Omnimaga founder
    • View Profile
    • Dream of Omnimaga Music
Re: Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #160 on: February 14, 2013, 02:58:07 am »
On the topic of the 7 inch screens, will the calculator still fit in our pockets or pen case like a Nintendo DS or a TI-89?

Offline Nick

  • LV9 Veteran (Next: 1337)
  • *********
  • Posts: 1166
  • Rating: +161/-3
  • You just got omnom'd
    • View Profile
    • Nick Steen
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #161 on: February 14, 2013, 06:26:21 am »
On the topic of the 7 inch screens, will the calculator still fit in our pockets or pen case like a Nintendo DS or a TI-89?
I'm afraid not, 7" is huge (too huge for a calc imo, I'd rather go with max 5"), it's this big:

while 5" looks like this:

Offline Lancelot_du_lac

  • LV3 Member (Next: 100)
  • ***
  • Posts: 53
  • Rating: +3/-0
    • View Profile
    • LI-Systems
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #162 on: February 16, 2013, 02:08:20 pm »
Lionel Debroux> ok i admit there is things that i've din't done so far. But there is Open Source schematics about arm board like rasberry, panda board, ...

Just need deleted useless parts and remap schematic smaller.

There is other board more interesting, ...

-------------------------------

About the look, i want do 800x480 or better, with a folding keyboard.

Offline DJ Omnimaga

  • Clacualters are teh gr33t
  • CoT Emeritus
  • LV15 Omnimagician (Next: --)
  • *
  • Posts: 55943
  • Rating: +3154/-232
  • CodeWalrus founder & retired Omnimaga founder
    • View Profile
    • Dream of Omnimaga Music
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #163 on: February 17, 2013, 03:38:27 am »
On the topic of the 7 inch screens, will the calculator still fit in our pockets or pen case like a Nintendo DS or a TI-89?
I'm afraid not, 7" is huge (too huge for a calc imo, I'd rather go with max 5"), it's this big:

while 5" looks like this:

Hmm for some reasons 7 looks too big but 5 too small. What about 5.5 or 6? (if that exists)

Also


Samsung Galaxy note 275.0 will probably have a 50 inch screen.
« Last Edit: February 17, 2013, 03:53:40 am by DJ_O »

Offline Lancelot_du_lac

  • LV3 Member (Next: 100)
  • ***
  • Posts: 53
  • Rating: +3/-0
    • View Profile
    • LI-Systems
Re: [new_calc] New Calc project ... with cas engine /!\
« Reply #164 on: February 17, 2013, 05:01:11 am »
not easy too found, so high price