calc84maniac wrote:
cvsoft wrote:
Edit: TI should focus on making a faster TI-68k model. The nSpire is slowly being forgotten, and 68k devices have more potential than TI might see. If they toss the V200's screen onto the TI-89 and add a backlight, they'd have one of my dream calcs. And from a programming view, the lack of the modulo function and the menu system on the 84+ series makes math programming more difficult. Dialogs or the TI-73's SetMenu( make input a lot easier.
Voyage 400, anybody?
I see your large screen and raise to you the PLT SU1.
Yup, it indeed seems clear that the Nspire series has replaced the TI-68k series. The 68000 and the later member of the series are good processors with a powerful instruction set, but ARM processors can be clocked much, much higher while remaining quite power efficient.
KermMartian wrote:
Sadly, the biggest lag is the tangle of code that is the equation parser. I'm actually continually surprised that it runs as fast as it does, after seeing from my work with AHelper just how much code that takes.
Are there, or will there be any info published? I'm kind of curious about it myself. It'd be fun to see just how much more convoulted MathPrint makes it, too.
Quote:
Quote:
The rechargeable battery is a nice inclusion for the average user, but limits the lifetime of the device to a few years before a new battery needs to be bought. Considering TI-81s are still in use, the 84+C won't get the longevity praise older models once got. People don't like having to pay for new hardware anyways.
I was just thinking that today. It'll be interesting to see in a year or a year and a half what students and teachers are saying about battery fatigue. Although I guess they've been dealing with that with the Nspire, and I haven't really heard too much complaining.
This is a concern I always tend to have with such devices (especially when the battery is internal and not user-replaceable). I've had relatively limited experience with Li-Ion battery packs, but despite what I've read about their shelf life, they frequently seem to have useful lives much longer than this info would suggest (especially the higher-capacity ones with a lot of extra “reserve” to begin with). I've seen some as old as 7–8 years still deliver reasonable runtimes, and I currently use a MiniDV camcorder of about that age with original batteries (which spent several years of almost complete non-use before I acquired it) which now have some modest capacity loss, but still deliver usable runtimes.
But no battery lasts forever, which is why AA/AAA/C/D form-factor cells are preferable, if not as convenient in ways. It seems “modern” Li-Ion stuff is too proprietary and expensive. Likely by the time I need to replace them, they won't be available anymore (not even for outrageously high prices) anyway.
Quote:
Quote:
Edit: TI should focus on making a faster TI-68k model. The nSpire is slowly being forgotten, and 68k devices have more potential than TI might see. If they toss the V200's screen onto the TI-89 and add a backlight, they'd have one of my dream calcs. And from a programming view, the lack of the modulo function and the menu system on the 84+ series makes math programming more difficult. Dialogs or the TI-73's SetMenu( make input a lot easier.
That would be amazing, but I think it's a pipe dream. As far as I can tell, the TI-Nspire CX CAS has more or less entirely ended the life of the 68k line. Here's hoping TI proves me wrong!
Well, I think the 68K calcs are still sold and in production, at least the TI-89t (this could have changed last time I checked, which would be quite disappointing, though inevitable for any product). Of course, it's obvious that no real development has happened with AMS in some time, and there aren't that many programmers for them these days nor much activity in the community (though that could theoretically be changed if enough people had interest).
KermMartian wrote:
Sadly, the biggest lag is the tangle of code that is the equation parser. I'm actually continually surprised that it runs as fast as it does, after seeing from my work with AHelper just how much code that takes.
Yup, the equation parser as far as I have seen has just been getting slower over time with the ti84+__ calcs... Also, the parser code for gCAS2 is indeed a frightful sight as is. I can barely read through it and understand all parts of it, but the complexity is needed to properly handle the math. I wonder if I can do a speed test of gCAS2's parser and evaluator against TIOS...
Travis wrote:
KermMartian wrote:
Sadly, the biggest lag is the tangle of code that is the equation parser. I'm actually continually surprised that it runs as fast as it does, after seeing from my work with AHelper just how much code that takes.
Are there, or will there be any info published? I'm kind of curious about it myself. It'd be fun to see just how much more convoulted MathPrint makes it, too. I could try to draw up some kind of high-level flowchart of how it works one of these days, although I'm not sure I'm the most qualified of our experts to do so.
Quote:
I've had relatively limited experience with Li-Ion battery packs, but despite what I've read about their shelf life, they frequently seem to have useful lives much longer than this info would suggest (especially the higher-capacity ones with a lot of extra “reserve” to begin with).
I'm more worried about what will happen to battery life as students and teachers use the calculators every day in math class, re-charging them every few days and then running them down for a few hundred cycles. I know laptop batteries tend to lose half their capacity in 200 or 300 cycles, so it seems to me the TI-84+CSE's battery will fall prey to the same phenomenon.
*bump* I just want to say that I read over the arguments in the ClrHome (or TakeFlight, I guess?) link in flyingfisch's signature, and I find the arguments to be extremely less convincing than my editorial "
Casio Prizm: Why TI Calc Coders Should Abandon the TI-Nspire CX". I'm afraid that rather than attacking the issue from an ideological or pedagogical standpoint (ie, "the TI-84+CSE will shoot curious students exploring programming in the foot") it reads like a promotion for the Prizm ("You should buy a Prizm because it's more powerful"). I will certainly agree that the Prizm is the more powerful, cheaper calculator, and that its more powerful hardware makes it more attractive for casual (C/Lua) programming. However, because the TI-84 Plus C Silver Edition supports TI-BASIC and ASM programming, and TI has made no efforts to lock us out of native programming for the device, I find articles claiming that the TI-84+CSE is bad for general programming to be disingenuous and misleading.
TakeFlight, actually. ClrHome is just offering hosting for it. One thing to note about TakeFlight is that the author comes from the Casio community so he might have a very strong bias towards Casio or against TI calcs
In my case you won't see a bad article or total bashing of the 84+CSE unless TI decided to remove ASM entirely or tried to hinder 3rd-party dev again. Granted, the speed isn't as great, but it will be entertaining to overcome the issues it create.
DJ_O wrote:
TakeFlight, actually. ClrHome is just offering hosting for it. One thing to note about TakeFlight is that the author comes from the Casio community so he might have a very strong bias towards Casio or against TI calcs
In my case you won't see a bad article or total bashing of the 84+CSE unless TI decided to remove ASM entirely or tried to hinder 3rd-party dev again. Granted, the speed isn't as great, but it will be entertaining to overcome the issues it create.
Agreed. I will certainly be very unhappy if TI tries to hinder native ASM or even Apps in any way (such as going back to requiring Apps to be signed/approved by TI). I'm glad to hear you agree that an "official" bashing of the TI-84+CSE by any of the community's leaders would be a poor decision at this point. Not that I have any investment in the calculator specifically, other than having fun with it, but I'm invested in TI's calculators, especially the TI-83+/TI-84+ line, continuing to be increasingly-featureful tools to explore programming (and math) for decades to come.
i realised just now that i never tossed in an opinion here.
from the perspectives of a programmer, a community member, and even a typical user, i am not looking forward to the CSE's release.
the programming complaint is fairly obvious. with an lcd that takes even longer to update than its already sluggish predecessor, combined with the impossibility of a screen buffer, i just can't see this model being useful for much of anything.
the community member position stems from this a bit. with the "shiny new" colour screen, the CSE is likely to steal away a good chunk of the userbase which would otherwise be using older models, putting them on a platform that's useless for programming and, thus, reducing the size of the community's future userbase.
the general user argument is one that people have been dancing around a bit with all the discussions of fancy-pants, toca la pantalla thingies that people seem to like so much these days. the only reason that a user (other than a student) would use a graphing calculator is an ability to perform mathematical operations rapidly and with minimal effort. the minimalistic interface of the 83+/84+ provides just that, with a practically instant boot time, no annoying gui to get in the way, and rapid access to most functions (including program creation) in just a few keystrokes. if i want a full mathematics suite, with clear graphs and charts, algebra and calculus, arithmetic on large collections of numbers, etc, then the obvious choice would be one of the many such programs that exist for a full desktop system. what's the point of getting a calculator if it takes several seconds to start and just as much interface navigation in order to start working? trying to shove such a system into a graphing calculator will always have fewer features and decreased performance compared to a full desktop counterpart (plenty of which are completely FOSS to boot), thus rendering itself obsolete. obviously, the CSE isn't taking it to this extreme, but it does present a decrease in performance (due both to hardware necessities and the inclusion of mathprint) that makes it a much less attractive option.
shmibs wrote:
i realised just now that i never tossed in an opinion here.
from the perspectives of a programmer, a community member, and even a typical user, i am not looking forward to the CSE's release.
the programming complaint is fairly obvious. with an lcd that takes even longer to update than its already sluggish predecessor, combined with the impossibility of a screen buffer, i just can't see this model being useful for much of anything.
The non-existence of an LCD buffer (in RAM) isn't the handicap you might imagine. We just got used to being able to design programs that could render out full frames in RAM, then use a relatively trivial amount of CPU time to copy those frames out to the LCD. Direct LCD writing is just another way to do things, and while it will require a new set of "helper" routines from shells (a la iFastCopy) and some different thinking from programmers, I don't think it's any better or worse than using an LCD buffer in RAM.
Quote:
the community member position stems from this a bit. with the "shiny new" colour screen, the CSE is likely to steal away a good chunk of the userbase which would otherwise be using older models, putting them on a platform that's useless for programming and, thus, reducing the size of the community's future userbase.
I fail to see where the "useless for programming" claim comes from in light of existing demos that show it's far from sluggish (for non-fullscreen drawing). You know I respect you greatly as a programmer, so I hope you'll clarify that claim for me.
Quote:
the general user argument is one that people have been dancing around a bit with all the discussions of fancy-pants, toca la pantalla thingies that people seem to like so much these days. the only reason that a user (other than a student) would use a graphing calculator is an ability to perform mathematical operations rapidly and with minimal effort. the minimalistic interface of the 83+/84+ provides just that, with a practically instant boot time, no annoying gui to get in the way, and rapid access to most functions (including program creation) in just a few keystrokes. if i want a full mathematics suite, with clear graphs and charts, algebra and calculus, arithmetic on large collections of numbers, etc, then the obvious choice would be one of the many such programs that exist for a full desktop system. what's the point of getting a calculator if it takes several seconds to start and just as much interface navigation in order to start working?
But it isn't. It starts as instantly as the TI-84 Plus, all of the menus are keystroke-by-keystroke identical, and there are only a few new Wizards over the previous MP operating systems. They really basically took the existing OS and just stretched the menus to 9 lines instead of 7 and threw in a few colors. Any argument in favor of the TI-84 Plus as a math tool applies to the TI-84 Plus C Silver Edition, and any con argument against the TI-84 Plus C Silver Edition as a math tool applies to the TI-84 Plus, in my opinion. The only difference is that you can tell multiple graphed lines apart thanks to the color, and you can put a nice grid behind things.
shmibs wrote:
i realised just now that i never tossed in an opinion here.
from the perspectives of a programmer, a community member, and even a typical user, i am not looking forward to the CSE's release.
the programming complaint is fairly obvious. with an lcd that takes even longer to update than its already sluggish predecessor, combined with the impossibility of a screen buffer, i just can't see this model being useful for much of anything.
the community member position stems from this a bit. with the "shiny new" colour screen, the CSE is likely to steal away a good chunk of the userbase which would otherwise be using older models, putting them on a platform that's useless for programming and, thus, reducing the size of the community's future userbase.
I completely agree with all of this. I am especially worried about the stealing away of the regular 83+/84+ userbase. Will the regular 83+ and 84+ slowly fade out of use? I understood that this would happen eventually, but I at least hoped it would be phased out in favor of a calculator that's nice from a programming standpoint as well:
KermMartian wrote:
The non-existence of an LCD buffer (in RAM) isn't the handicap you might imagine. We just got used to being able to design programs that could render out full frames in RAM, then use a relatively trivial amount of CPU time to copy those frames out to the LCD. Direct LCD writing is just another way to do things, and while it will require a new set of "helper" routines from shells (a la iFastCopy) and some different thinking from programmers, I don't think it's any better or worse than using an LCD buffer in RAM.
It's definitely worse. Games and utilities have the potential to look beautiful in color, but unless it's mostly a static base image with moving sprites on top of it, it's going to have to be a beautiful slideshow. As far as I know, it will be impossible to implement some things in a fast manner, like a smooth multi-directional scroller.
Psh, it won't be that difficult
For multi-directional, I think 10FPS is sufficient and plausible for animated tiles, maybe 5FPS to handle compressed sprites. I have been working on some graphics routines and they aren't that bad, actually. For example, circle drawing is about the same speed as if you didn't update the LCD.
"useless for programming" was definitely an overstatement. that should have been "useless for anything i, personally, would want to program". screen buffers may not be necessary for drawing, but they are much faster for things like full screen transformations and very convenient for things like the layering of multiple images. of course, these two things are no longer relevant anyways if the maximum fullscreen refresh rate (that is, updating every pixel on the screen at once) is 3-4 fps (i'm not sure exactly where i heard that number, though, so feel free to correct me.) the CSE introduces a large, full-colour screen and then denies me the possibility of creating any of the things i would want to make to take advantage of it.
as for the math usability issue: the "several second boot time, guis getting in the way, etc" complaint was more directed towards the nspire and prizm. what i was trying to say was that any decrease in speed, which the CSE supposedly presents, is a step backwards in functionality.
On the other hand, drawing sprites is faster, easier, and the code is smaller (though the sprites are much larger if you use the full color).
First, think about this: Grayscale on existing calcs requires a large amount of CPU time in order to send bytes to the LCD near 60Hz. With the TI84PCSE, this isn't the case any more. You can easily stream the sprite data once and the CPU will be free to do other things.
Also, I hope TI makes more variations of the calc so I can string together longer names like TI84PCSEPFR (Imagine `Texas Instruments 84 Plus Color Silver Edition Pocket French' on a package)
I'm not concerned about the loss of 84+ popularity in short terms, especially if they drop the price of the TI-84 Plus, but in long terms we will need a parser that can run old games, even if a bit slower or even if not full screen, so that 84+ users won't lose much audience (kinda like when the TI-Nspire Clickpad came out with the TI-84 Plus keypad).
Performance-wise, I am worried about animated tilemaps and such intensive games, but for other games I'm not, since all it takes to get around the limitations is creativity when coding. We did it with the TI-82, we did it with the TI-83 Plus, so we can possibly do it with the TI-84 Plus C Silver Edition.
What helps a lot now is that since the new calc has ASM that is similar to older models, we already have a lot of TI-84 Plus documentation and tricks out there, so it won't take as long for ASM coders to get used to the platform. It's made even easier now with the fact that a community emulator will already be out by the time the calc comes out, which wasn't the case with the TI-84 Plus.
For BASIC programmers, some might say that they don't care and they should just switch to ASM, but they'll still be able to get around limitations if they want to code properly. I already have some ideas to test when the calc comes out, and if they all fail, then there's always good ol' ASCII art awaiting for me (assuming I code).
By the way, two plus that the calc got over two other models:
-Instantaneous boot time: On the TI-Nspire, it takes up to 1 minute if in standby mode or after removing batteries. On the PRIZM, it takes several seconds of setup.
-Instantaneous turning OFF: On the Casio PRIZM, it takes a few seconds to turn OFF. On the Nspire it does as well if you got an high score in a Lua add-in.
I haven't gotten a chance to use one, but I think the 84+CSE looks pretty cool. From what I've heard, the programming seems just fine for a more casual programmer like me, and the color screen is pretty awesome. The 21K of ram is really disappointing, though, but it does have almost 3 times the Flash ROM as the 83+. However, I do totally agree with cvsoft:
cvsoft wrote:
TI should focus on making a faster TI-68k model. The nSpire is slowly being forgotten, and 68k devices have more potential than TI might see. If they toss the V200's screen onto the TI-89 and add a backlight, they'd have one of my dream calcs
Its a z80 calculator with a colour screen, how can that be bad?
tr1p1ea wrote:
Its a z80 calculator with a colour screen, how can that be bad?
That's pretty much my feeling as well. I think it appeals to at least three groups:
1) Math students and teachers - color must mean it's more powerful, right? And the color really does help make graphs clearer.
2) Beginner TI-BASIC coders: they won't know what they're missing
3) Advanced ASM coders: they'll enjoy the challenge of a new device bringing a new set of limitations.
As far as I can see, there's really only one (lumped) group of people to whom it might not necessarily appeal: intermediate/advanced TI-BASIC-only coders, and Axe coders. They'll need to adapt the most to the new device.
KermMartian wrote:
tr1p1ea wrote:
Its a z80 calculator with a colour screen, how can that be bad?
That's pretty much my feeling as well. I think it appeals to at least three groups:
1) Math students and teachers - color must mean it's more powerful, right? And the color really does help make graphs clearer.
2) Beginner TI-BASIC coders: they won't know what they're missing
3) Advanced ASM coders: they'll enjoy the challenge of a new device bringing a new set of limitations.
As far as I can see, there's really only one (lumped) group of people to whom it might not necessarily appeal: intermediate/advanced TI-BASIC-only coders, and Axe coders. They'll need to adapt the most to the new device.
I would call myself an advanced assembly coder, and I love a challenge. Challenge me to make a smooth multi-directional scrolling tilemap that runs at 60FPS on an 83+ and I'd love to try it. But challenge me to do the same on an 84+CSE and I'll just tell you it's not even close to possible, because it physically isn't. No optimization can make it nearly fast enough. Working on a system with these kind of limitations isn't a fun challenge, it's an aggravating challenge, because you know that no matter what you do, you'll never be able to do some of the things you could do on an 83+.
I recognize that perhaps this sort of comes with the territory of working with a z80 and a 320x240 16-bit color LCD, but the fact that TI made absolutely no efforts to assist faster graphics pisses me off. Memory-mapped buffer/sprites? Nope. LCD paletting? Nope. Driver-powered upscaling? Nope. Faster processor? Nope.