...what would it be? Are you curious about something with their past, present, or existing product lines? Are you curious why Lua, why z80/ARM, why not 68k? Curious what activities are next? What the future of TI calculators and the SATs are? I have no idea if any of those questions have publicly-available answers, but I encourage you to tell me all the burning questions you have about TI's calculators below. I have only two requirements:
1. Any post that asks a variation on "Why are they so expensive?" will be ignored. I already have a diplomatic version of this question written down.
2. Technical details like "where are the missing USB documents?" or "how do we control the charging LED?" aren't the sort of things I am looking for.

So, get to it! Lively, respectful debate/refinement of questions is also welcome.

Why was the processing power of the 84+CSE not improved along with the LCD

Why do TI somewhat shun the community, despite us actively embracing the principles of STEM which they so actively promote?

Is there still a chance that the 68k line will be improved?

Why, on the TI84PCSE page, does it say there's an app SDK available and there's no download link?

Good questions so far. I know the practical answers to those (in reverse order, because they didn't finish making it yet, no, they don't any more quite as much, and because that was the least expensive option), but it's possible there are better answers. What else?

Based on our overclocking experiments, it seems that 25 MHz might not be attainable, which was TI original estimate for the maximum clock speed of the TI-83+SE (which was a sort of beta version of the TI-84+ series). Furthermore, the specified accuracy of the CPU clock frequency is only 20 %. They might have boosted the specified CPU speed to 20 MHz, but +20 % (24 MHz) would have put them over our empirically determined maximum of 22 to 23 MHz. So they would have had to switch to much more accurate resistors and capacitors for CPU clock generation, and only gotten 5 more megahertz. Also, it would require additional modification of the assembly line. While the amortized cost might be less than a dollar per unit, the cost of additional testing to verify that all their parts would work at the higher speed might be thousands of dollars. Or, maybe they did do the testing and found that some percentage of ASICs wouldn't even operate at 20 MHz.

Alternatively, they could die shrink the ASIC, which might permit much higher frequencies. However, there's no guarantee that might work, their license to produce embedded Z80s might not allow that, and they'd have to move production to a completely different foundry and produce new die masks, which are two very expensive things to do. They might anticipate only a few hundred thousand TI-84+CSEs being produced, leading to a cost of several extra dollars per unit for the extra speed. You might say it's worth it, but TI admits to being in the education business solely because it's very profitable; education is actually only a very minor contributor to the company's overall revenue. So TI doesn't see it as a worthwhile investment.

TL;DR: Their current production process doesn't allow higher clock speeds, and it would might cost tens of thousands of dollars to upgrade their production line, which they don't think is worth it.

It's worth pointing out that we're talking about TI here, who own foundries at many different semiconductor process nodes. I think the bulk of the expense would be fabbing and testing small-batch copies of any new ASIC design, and doing the necessary setup to produce millions of the new design. Not to mention that they probably have hundreds of thousands of the "old" design saved up.

Will the bugs reported over a year ago in the TI-84 Plus C SIlver Edition OSes and the TI-Nspire OS be fixed in the future?

I already asked about the TI-84+CSE bugs (although it's always worth asking again); you can find their answers here: http://www.cemetech.net/news.php?id=660 . To which TI-Nspire OS bugs do you refer?

What does TI see as the next design of calculators in their product line (as in, not an upgrade of a previous model, but truly a new device)?

How does TI see their devices impacting education in departments other than math and basic science?

What is TI looking for from the community? How can we promote a friendlier, open-minded relationship between enthusiast community and big corperation?

What would TI's reaction be if a z80 CAS were made?

Why do you refuse to support Linux?

Now that the 84PCSE has been relatively well received besides the ongoing complaints about responsiveness, do they have any plans to release a higher-end model with an ez80?


I appreciate their new-found (or newly-rekindled) willingness to interact with the community, but this particular issue seems to be a stupid oversight on their part. Not only do they have a community that is willing to report bugs, they're getting free labor in the form of patches from people like Brandon, and they're not taking advantage of it. At the very least adding a real bug-reporting mechanism, even if it's limited to verified/trusted submitters to cut down on noise seems like it should be valuable for both sides.




Finally, any (long-term) plans to support complex matrices and/or custom matrix names (in the same way we can currently do for Lists)? Both of these would go a long way to making the matrix functionality a LOT more useful to physics users and programs which need to do a lot of vector operations.

I would like to ask them to make their calculators open to development completely, and if they could provide in-depth manuals/schematics of the calculators and support this open development.

Excellent question.

I completely agree. We not only have the expertise and commitment to find the bugs that even QA misses, some of us have the expertise to figure out what the fixes would be. I think that's definitely a relationship that could help both us and TI.

It would be nice to know why they broke the menu equates.

I would also like to know why they don't optimize more. I've shown that careful optimization in display routines alone can make a textual user interface that's much snappier than the OS (which is, face it, textual). CabriJr and CelSheet seem particularly slow, and they're far more graphical. In CabriJr, it seems like it takes a full second for the cursor to respond to each keypress! (I've got a resume to send them, if they're interested.)

I agree that there's a lot of graphical optimization that could be done; I feel like those new Apps definitely show the fact that they've been ported from the monochrome versions with minimal modification. I think perhaps the more general question is whether they are willing to have engineers that will continue to work on new Apps as a dedicated job, which would mean said engineers would be able to learn more techniques and strategies for writing tight, fast graphical code for the TI-84+CSE's LCD. Unless the author of those Apps is Pat Milheron, in which case I take everything back.

*bump* Any more feedback, questions, comments, or vitriol for this topic? Speak tonight or forever hold your peace.

I also think that this could be really nice if implemented with pics.