hp graphics - comment?

[lokhin]

hello

not long ago, I started a thread about hp 28s. both jeremy, john who comment on its graphic abilities as weak. and other dont use it at all.

I read other reviews abot newer hp graphics and most reviews commented that hp graphics perform better in number chruching and weaker in graphing,even in newer model -hp 48g they said that the graph plots and much slower and zooming is more difficult than TIs. is graphing an hp weakness? what are their strength then?

seller retracted the auction I mentioned last time... no chance to get it... but the calcs that I cant get is bad, I believe.

-lokhin

[Ellis Easley]

The TI's come with different pixel resolutions. I have a TI83+SE which has 96 X 64. The 48's are 131 X 64. I think the higher end TI's have higher resolution than the 48's. More pixels take more time to draw (all other things being equal). This might have some bearing on the relative speed of the plotting functions. Another important factor might be the precision used to calculate the pixel coordinates. Since there are less than 256 pixels on either axis, 8 bit integer math would be sufficient. I don't think the processor in the 48's is capable of low precision math. The TI83 has a Z-80 processor and even if it is emulating high precision BCD math to evaluate the function to be plotted, it can use 8 bit integer math to calculate the physical pixel coordinates once the function value is calculated. Scaling and clipping are two steps that have to be performed for each pixel after the function is evaluated. If the TI's are doing these steps with 8 bit integer math, it would take much less time than the full precision BCD math that the 48 has to do. One final thing that might be involved is the range checking that the 48 operating system does. As I understand it, the difference between User and System RPL has to do with this range checking that is done before every operation in User RPL, but which is bypassed in System RPL. I wonder if there is any way to specify the equation on a 48 using Sytem RPL?

[Vieira, Luiz C. (Brazil)]

Hi;

I am not sure about it, but I think it's possible to find integer numbers faster than a full-precision floating point in the same expression. If so, I think it would be wise using an integer-number ploter program to evaluate the expression and give final coordinates directly. I am beguining to use a TI82 (thanks to Michel Beaulieu) and I see it creates faster table and ploted dots. I see that resolution interferes in table-building speed.

Maybe thsis is the goal: why to obtain floating point numbers from an expression if I need only integers to plot it, say, to plot pixels?

Just a guess.

Luiz C. Vieira - Brazil

[hugh]

the 28 has such a small graphic screen it is silly and the 48 has all the features but is just too darn slow. embarrassingly slow, in fact.
i am sorry to say this because i got excited by both machines initially because they could actually do graphics at all. nowdays its commonplace.

[Speck]

The TI's are a bit different from machine to machine. The 91/92+/V200 (being the same calc) use the same screen. I don't remember the exact numbers, but it's about 220x120 pixels. This makes it great for graphing; less so for toting around. I believe 82/85/86 use pixel resolution similar to that of the HP-48/49, but the HP utilizes screen real estate arguably better. I believe the TI-89 is somewhere between the 48/49 and the TI-92 (they had to sacrifice a few pixels here and there to make it all fit). Graphing will be task dependent. 85's have zipped past my 92 on a couple of occasions, and my 92 has zipped past PC-based programs several times, all dependent on what was being graphed. I don't have a lot of experience with graphing on the 49G, but as long as it has that capability and is easy to carry around, I'm happy.