toma@tekchips.UUCP (Tom Almy) (02/09/84)
I have such an investment in fig and 79 code (which are nearly compatible) that I have no intension of going to the 83 standard! EVER! The difference in tick ( ' ) is just the start. Division works differently, PICK and ROLL have changed (the worst disaster of all!). In fact, compatibility with previous Forths was a low priority of the standards committee. To make matters worse, there is a 83 standard "model" that is being distributed, public domain, that has a great many extensions to th standard. This model will undoubtedly become the defacto standard! Programs written to the model will probably not run on an arbitrary standard system -- in fact, none of the utility programs provided with the model would run in an arbitrary 83 standard system! The 83 model also fails to define the vocabulary mechanism. The fig scheme, 79 Standard, ONLY...ALSO (used by the 83 model), and STOIC vocabulary stacks would all meet the standard -- yet you cannot have code portability between any of these unless you use vocabularies in the most trivial ways. The 83 model is also only applicable to 16 bit integer machines, making it essentually useless for VAXen, 68ks, 16032s, and what ever the future holds. I have made good use of Forth on VAX and 68000; it is not a toy language! Yet the standard won't let the language grow. (Maybe I should have posted this to net.forth.flame!) Tom Almy, Tektronix Inc.