Compilers vs. Interpreters

----- Original Message ----- From: "Bob" <zzzzz at sbcglobal.netTo: "General Discussion: On-Topic and Off-Topic Posts" <cctalk at classiccmp.orgSent: Monday, October 23, 2006 2:28 PM Subject: Re: Compilers vs. Interpreters modifiable code enviroment. Many years ago, I programmed in a language called CREDIT, running on Philips P6000 series. This more-or-less interpreted language had some remarkable features. It had a kernel handling the (1 byte) opcodes. For data, 16 blocks (IIRC) were available, but redefinition was allowed. For each data item, max 16 per block, a table was kept specifying the peculiarities of each item. So, a simple MOVE would demand 3 bytes : 1 for the opcode, 1 for the source and 1 for the destination. A MOVE between two indexed items would take 5 bytes, as two bytes were needed for the indexes (1 byte each). This would make it extremely easy to modify the instructions, as you didnt need to know much about the data. Also, the resulting code was so compact, that we could accommodate a back-office routine for entering financial data (30-40 different record types), plus a routine for transmitting the data via a 3270 connection, within 32K bytes Nico