# HP Forums

You're currently viewing a stripped down version of our content. View the full version with proper formatting.

Ok, so I bought a 50g to play with, and I guess in a way to finish what I started when I bought my 28s so long ago. So here I am with loads of RPL info. I printed the entire 50g User's Guide, the first 120 pages of the 49G+ AUM (which covers RPL). I also have the 49g AUR and the Programming in System RPL book (which I assume will work for the 50g since it also covers the 49G). All together about 2,000 pages of information, much of which I assume will be redundant.

I wasn't looking for a new career, just trying to finally do something I had tried and failed at in the past (mostly due to a lack of motivation). But my question is this: How much of all this stuff really necessary? I feel more than a bit overwhelmed.

Greetings David! I can certainly see where going from a 28s to 50g would be a shocker! In one step you are skipping more than 15 years of HP calculator innovation (and stagnation if you include the dark years of Kinpo).

May I recommend learning the 48g before FULLY tackling the 50g? I believe this will bridge the RPL complexity gap as you follow the core RPL components through their natural evolution.

The most complete calculator in each family is 28s -> 48g -> 50g, but there is not a natural jump straight from 28 to 50.
(I am purposely skipping the 48s, 48sx, 48g+,48gx, 49g, and 49g+ in the chronology and maturity because by analogy:

1. the 49g/49g+ is to the 50g what the 28c is to the 28s,

2. also, the 48s is to 48g what the 28c was to the 28s,

3. you really don't need the extra expense of the 48SX/48g+/48gx to learn the RPL language for the 48 series calculator)

Others' opinions may differ, but I started RPL with the 48g and found it very easy to port the lessons learned both ways to the 28s and the 50g. If I had tried only both extrema, I fear I may have given up.

Edited: 16 Aug 2007, 12:03 a.m.

http://www.hpcalc.org/details.php?id=1771

This is a nice little tutorial about programming. You'll learn more about the system and using it by doing these examples than you'd ever get by reading manuals. I ran through this tutorial back in . . . i think it must have been 8th grade. Later moved on to sysRPL and never really looked back.

TW

No reason to master everything on day one.

I'd seriously punt on the SYS RPL stuff until you REALLY think you need it. User RPL is dangerous enough as it is.

When writing code, write lots of little functions. They're easy to debug and test. Just place a few bits on the stack, and work it until it's done. Then build up from there. You don't have anywhere near the memory constraints on these machines as you do on the normal calculators. They pretty much support "unlimited" everything.

Also, they have a really nice single stepping trace/debug mode built in. That's a wonder.

As with anything, just work on whatever application or utility you want rather than trying to use all of the features.

Developing on these machines I found very organic, and I never used a PC or anything to work on them -- just for transferring files. Some folks use tools to use PCs to edit or even use as a terminal to these things. I never bothered, easier to just flump on the couch and code away. But I never did any SYS RPL or assembly stuff. I can easily see folks doing that on a PC.

So, work on your application that you want. No reason to master it all -- just what you need.

That's the best way to get started. Unless you code every week with it, you'll forget some things, but like bike riding, you can't forget the basic approaches and structure. And this is where the AUR is most useful, as a reference guide for commands and structure.

As far as programmin in RPL the 50G is almost the same as the 48G/GX. And not too far from the 28C. I think you can pretty much run UserRPL programs for the 28c on the 50G.

Quote:
Greetings David! I can certainly see where going from a 28s to 50g would be a shocker! In one step you are skipping more than 15 years of HP calculator innovation (and stagnation if you include the dark years of Kinpo).

May I recommend learning the 48g before FULLY tackling the 50g? I believe this will bridge the RPL complexity gap as you follow the core RPL components through their natural evolution.

The most complete calculator in each family is 28s -> 48g -> 50g, but there is not a natural jump straight from 28 to 50.
(I am purposely skipping the 48s, 48sx, 48g+,48gx, 49g, and 49g+ in the chronology and maturity because by analogy:

I am doing mostly financial stuff, so I have just purchased a 12c (longing for my original 11c, but, well, financial), and a 19b II (same problem.) I sold the 28S a long time ago, but nostagalia is starting to get expensive!

Anyway, if I am understanding you correctly I should go get a 48G to learn RPL before I try to tackle it on my 50G? It looks like I could get on at a reasonable price, and I just bought a printer that has auto duplex, so the manuals wouldn't be a problem.

Thanks to all for the doc links, I will look at them as well.

Quote:
Anyway, if I am understanding you correctly I should go get a 48G to learn RPL before I try to tackle it on my 50G?

That would be my recommendation because of it's centrality between your newest RPL and the one you learned back in the day. Again, others may have a different opinion. You can also use the integral TVM (TVMROOT) features of the 48g if your 12c is out of reach.. just bear in mind the treatment of annual/periodic interest rates are different for each HP model.

Best of luck in your financial project.. unfortunately the best financial calc (17bii) is neither RPL nor programmable. :(

Quote:
How much of all this stuff really necessary?

These calculators are powerful enough (especially the 50g), that I am comfortable saying that these are solutions looking for problems. You can get into as little or as much of the RPL as you like. Most of it I find completely worthless. Really! HP advertises over 2300 built-in functions. Compare with studies in Lexicography (link has good references for numbers), that show one can have 75% coverage of an entire written language by using only 1000 words. (By contrast, the venerable 42s offered around 600 functions.)

Any thoughts from the forum on applying the Nation/Waring conclusions to deciding how many RPL 'words' are sufficient?? GRIN.

Edited: 17 Aug 2007, 12:55 a.m.

Fascinating!

I was under the impression that one needed more words in one's vocabulary to have the "coverage" of the language that the paper(?)in the link listed.

However, there are functional differences between a real, human language and a computer "language", which really is the use of words from one's language to name operations and conditions strictly for use in defining, broadly, algorithms.

As such, it should require "only so many" "words" to define any set of similar algorithms. Now, should a user need to define more algorithm types, then he needs more or new words.

Therefore, perhaps "most people" may only need to know a small set of RPL commands or functions, but as one needs to apply his 28S, 48S, 48G, 49G, or 50G to more types of problems or uses, then his need for a more comprehensive command of RPL grows, as well.

I mean, the occasional need by a scientist would require less of a knowledge of RPL than perhaps the more extensive need of a technical student to solve homework problems, etc.; but which differs from the more intensive need maybe of the engineer who may have a less wide but deeper set of applications.

Oh... don't let this out to kids in school; they're lazy enough as it is. If word gets out they only need x number of words to master English or Spanish or German 80%, then I suspect they will begin to study only 80% or less as hard as they might've otherwise! ;)

Quote:
unfortunately the best financial calc (17bii) is neither RPL nor programmable. :(

Oh, Hudendai! See Article 712. I wouldn't call the solver the world's best programming platform, but it does have some features that make it *minimally* acceptable. This is especially true for the original 17bii. The + solver has some issues.

Quote:
I wouldn't call the solver the world's best programming platform,

I agree 100%.

I think it is the best financial calc, and explicitly not programmable- even the solver tricks aren't what I would call minimally acceptable- at least not for programming. A good example: if you try an amortization in the (all?) pioneer solver, the result is MUCH slower than the 12c. Should I need to program anything, it is easier to go to another (RPL) platform or 42s.