home *** CD-ROM | disk | FTP | other *** search
Text File | 1990-02-23 | 1.7 KB | 40 lines | [TEXT/GEOL] |
- Item forwarded by A33 to A34
-
- Item 9745731 20-Feb-90 14:47PST
-
- From: PASCOE1 Pascoe, Geoff
-
- To: MACAPP.TECH$ MacApp Technical
-
- Sub: Persistent about Persistence
-
- Greg, et al,
-
- I think it would be a mistake to overload Pascal (and MacApp) with lots of
- useful, but very complex, run-time and language semantics. For example,
- although a virtual object management system is useful, I think a lot of users
- would feel that cost and complexity would outwiegh the benefits (reread Greg's
- latest link on this subject- shared locks, exclusive locks, LRU algorithms-
- man, this is turning into an operating system). A big issue here is how far
- can you take Pascal before it's not Pascal anymore. If you really want this
- stuff, maybe Lisp or Smalltalk might be better lanugages for you.
-
- Now, let me qualify my previous statement by saying that Object Pascal can
- stand to have a little more run-time without losing it's essential character.
- I suggest that complete metadata (this is mostly a compiler issue) and run-time
- for a good automatic I/O is about as far as we should take it. A 'perform'
- facility would also be nice.
-
- As far as C++ is concerned, I don't think its the answer at all. The
- overloading of -> is probably one of the smaller issues. Consider that C++,
- overall is much more complex than Object Pascal and it's run-time is
- considerably more impoverished.
-
- Geoff
-
- P.S. On the subject of O-O Virtual memory, I suggest you read about LOOM (Large
- Object-Oriented Memory for Smalltalk-80 systems). There's a paper by Kaehler
- and Krasner on this in the book "Smalltalk-80 Bits of History, Words of
- Advice". You might find the concepts useful in your work.
-
-