I know... I should have listened to you.

Here are JUST my math gripes from HP forums that I contributed to (so this doesn't even include games, interface, programming, etc.):

No, you can't specify log of 38 base 5 by default. There is probably a program for this, or you can easily make your own using the change-of- base formula. For example, to calculate the above log, just do log(38)/ log(5) or ln(38)/ln(5), but you probably already knew that. --Graywolf

Yes, I too was bitterly disappointed by the HP-50g's CAS. I think I should have stuck with the 89 a little longer... --Graywolf

Well, I had been using the 89 for three years compared to 3 months of the HP 50g. I find that the CAS on the 89 is *overall* better than the 50g. Let's see, here are some few points that I can remember off the top of my head:

1.  The symbolic solver on the 89 is WAY better than the HP. The HP is limited to rational expressions and elementary trancedental functions. The TI can adquetly handle polynomial expressions, logarithms (completely unlike HP), trigonometric functions (completely unlike HP), e and pi functions, and a few others I cannot remember. The point is, the 89 will try to solve everything you throw at it symbolically, and if it can't, it will give you a numerical answer. I find that if an equation has a symbolic answer, the 89 WILL be able to solve it and will not give up like the HP if it doesn't simplify to a rational expression.

2.  Symbolic integration (definite and indefinite) on 89 is better than HP. I can't say much more about that. There were a couple of integrals I had written down, and if I find them, I will post them.

3.  Simplification on the 89 is vastly better than that of the HP. I know this is a point of contention because some people want internal simplification whereas others do not. For example, entering sin(pi/2- x) onto the 89 will return cos(x) whereas the HP will leave it unchanged. Of course, you can go to the TRIG menu and rewrite it in many different ways. I find that this is burdensome and the CAS has no option to do this internally. Also, sometimes I expect a command to change everything into sin() or cos(), but it doesn't so I have to semi-randomly try functions to get an answer that I can write down on paper.

4.  The 89's CAS trys to be consistent (note I didn't say it was always) whereas the HP will constantly ask you to switch modes and not switch to the previous ones after a calculation. I think everyone knows this already and I'm sure there are programs for this, but I don't want to take the time to download them , I can do it manually (but it's still a pain).

5.  I find that the TI is better at calculating limits using the limit() command on the 89 and the lim() command on the HP. The series command is so-so, sometimes it evaluates a limit, otherwise it returns ? which I take means "undefined" or "I can't do it".

6.  This is contentious: the 89 is better at summation and products. I find that the HP sometimes returns summations as psi(a)-psi(B). This is good that the HP is making use of it's built-in commands, but for something like the sum of reciprocals, I sometimes want the fractional answer. And "arrow Q" function doesn't help for large fractions. And infinite summation is very weak on the HP. Also, I haven't been able to find a product command like the 89's capital PI command. (I haven't looked in the manual yet...)

7.  Factoring, generally, is better on the 89 because of speed. Nothing much to say here, it's not a major issue because both will probably handle whatever you throw at them.

8.  This might be simplification or maybe not, but it's your call: the 89 always tries to resolve everything in symbolics in Exact mode. The HP, on the other hand, will ask for mode switches. For example, the 89 can evaluate sin(pi/5) whereas the HP can't (symbolically).

This is all that pops into mind right now... All of the above is in my opinion so HP fans and TI fans, do not take offense. All I can say is that the 89 has a superior CAS mainly because it is a mini version of Derive. --Graywolf

OK. I'm sure the 89 can't do 120x120 equation solving that fast, but I think most people shouldn't be worried about that (because the average person hardly needs to solve over 4 unknowns). I know that the 50g hardware is superior to the 89 so arithmetic should be faster, but I also have heard that the CAS is emulated so it doesn't take full advatage of the speed. I have heard the HP is good for linear algebra and I admit it is true, but the 89 is not incompetent. The 89 just lacks roughly 1-15% of the benefits.

And the list isn't too long; it grows (and I'm not bashing HP here, I can say a lot of positive things about the 50g):

9.  The 89 has "smart" equating features (this is mildly contestable). For example, you can enter (x^2-4)/(x+2)=x-2 and the 89 will return true. On the HP, you would type the exact same thing, except it would be seperated by ==. The HP would return 1 (true). However, the HP breaks down very fast (I don't know why) because it will not equate sin(pi/2-x) with cos(x) as the 89 will. Furthermore, the HP lacks advanced absolute value and inequality support.

10. The HP has weaker complex numerical support. I don't have an 89 with me (stolen  ) so I can't give examples, but from the manual I still have I can tell that the HP has a difficult time integrating complex functions. This is also true with some of the other calculus functions such as TAYLOR.

It's nearly 10:00 PM here so I'm going to sleep now. I'll try to find more gaps in the HP CAS tomorrow. --Graywolf

I can only speak of the 50g. The HP Solve is NOT REALLY all that legendary. You can think of two solvers on the 50g: "normal" and differential.

The differential solver is very versatile (at least with my minimum use) and is reasonably competent for symbolic and numerical needs.

The "normal" solvers include: numerical solver, symbolic solver, linear solver, multiple equation solver. The numerical solver is good (like you would normally expect it to be), but nothing legendary. The symbolic solver just plain SUCKS; it is only restricted to polynomials and very elementary trancedental functions (and it excludes some). The linear solver is normal, nothing legendary. And the multiple equation solver is good.

So, the only good things, in my opinion (of course), are: lots of options (they could all be bunched into three programs personally), nice differential solvers, nice multiple equation solver.

Overall, I think it's so-so. Expect good numerical results, don't expect any symbolic feats. --Graywolf

 So, I bought an HP 50g recently and I have only had experiences with a TI-89 Titanium, never an HP calculator. So, naturally I have a few questions:

1. How do I solve a^3+a*b^2=30 and b^3+b*a^2=90? The answer is 3^(1/3) and 3*3^(1/3). The 89 can solve this easily.

2. How do I solve (n-23)!23!=(n-32)!32!, not graphically, but numerically for integer values greater than 32? The answer is 55 and the 89 can solve this using nSolve.

3. How do I solve 2007^(1/a)*2007^(1/B)=2007^(1/9) for a and b with n or c being constants? That is, how can I get a dependent solution where a and b will be some function of another variable. The 89 can do this.

4. How do I symbolically solve log(x_2)/log(2x_4)=log(4x_8)/log(8x_16)? The answers are 1/16 and 2. The 89 can do this easily.

 7. How do I test if equalities or inequalities are true or false? That is, (x^2-1)/(x-1)=x+1 (true or false?), sin(x)=cos(pi/2-x) (true or false?), 1>2 (true or false?). --Graywolf

And that's only some of the things I bothered to post online.

EDIT: No idea why some of the smiley faces are there.