Well, to do some of the math functions, there is almost no difference in speed (like if you were to make your own floating point math routines, it would be similar to the OS codes). However, in almost all cases the speed increase is huge due in part to the fact that we usually don't use floating point numbers on the calc. Plus, we can draw directly to the screen to make sprites instead of using pixels and lines and stuff...

Yeah... but I have been working on a code interpreter and interpreted code is about as fast as compiled code. It is actually all the parsing that slows BASIC down. Like for example, if you did sum({0,1,2,3,4,5} it has to convert all of those numbers that are separated by commas into 9 byte floating point numbers and it has to create a list temporarily that has those elements then it executes the sum( command. If they had made it "tokenize" the way it is done on an 89, the BASIC codes would be interpreted so much faster.

The downside is that tokenizing code for the first time may take a little while, but after that it no longer needs to tokenize (until the program is edited again). Also, tokenizing sometimes shrinks the program a little.

Isn't the speed advantage huge? Especially on 83+'s, which I usually use now when I borrow from people (I lost my nSpire a little while ago)

It depends on what you're doing. The OS itself is written in ASM, so if you were to try doing the same things that the OS has to do to run BASIC, there'd be no significant improvement.

In TI-BASIC? expr(Str1->A will convert a string into a number and store it in A. If you only want to convert part of the string, use the Sub( command as well.

In Axe, strings are already numbers. Just subtract _E30 from the value pointed to by whatever pointer you're using.

ah, that's fine greg

can't believe I forgot about Expr() though O.o