home *** CD-ROM | disk | FTP | other *** search
Text File | 1991-03-06 | 1.8 KB | 49 lines | [TEXT/GEOL] |
- Item forwarded by FRED.FORSMAN to DIEZMANN PZZA.SLT
-
- Item 2677448 19-Sept-90 05:33PDT
-
- From: POWERUP.ENG Power Up Software,PRT
-
- To: AUST0134 Jam Software Sydney,IVR
- MACAPP.TECH$ MacApp Technical
-
- Sub: RE>FWD>Jump table blues
-
- Attn: Jam Software Sydney,IVR
- Attn: MacApp.Tech$
- SentBy: James Plamondon
- Date 9/18/90
- Subject RE>FWD>Jump table blues
- From James Plamondon
- To Jam Software Sydney,IVR
- CC MacApp.Tech$
-
- Subject: RE>FWD>Jump table blues
- Dear Tseung,
-
- The ETO linker borrows space from the 32K segment allocated to hold globals,
- and uses that borrowed space to hold additional jump table entries. (This is
- second-hand information, and may be in error.) Thus, the ETO linker raises
- the limit on the number of jump table entries (assuming that your globals take
- up less than 32K, which is a safe bet in most OOP programs, I should think).
-
- However, the number of jump table entries is still bounded. This limitation
- is scary, because it raises a significant barrier to the adoption of OOP for
- the kind of project for which OOP is best suited — very large software
- projects. The combination of standard OOP practice (which suggests the
- writing of very small methods) and large software projects, results in the
- creation of zillions of little tiny methods, blowing the linker all to hell.
- (Then, there's the memory manager… but that's a different link.)
-
- I have enough respect for the Apple engineers to be confident that they are
- working on this problem. The ETO linker is evidence of this (although it is
- not the final solution). I would be delighted to hear the thoughts of someone
- involved with the MPW Linker project on this subject.
-
- Looking forward to an unchained linker, I remain
-
- Yours,
-
- James Plamondon
-
-