home *** CD-ROM | disk | FTP | other *** search
Text File | 1991-03-06 | 4.4 KB | 84 lines | [TEXT/GEOL] |
- Item 9511310 16-Feb-91 11:35PST
-
- From: ALGER Alger, Jeff,VCA
-
- To: MACAPP.TECH$ MacApp Technical
-
- ------------------------------------------------------------------------------
-
- Sub: Healing the MacApp/C++ rift
-
- Fellow MacAppers,
-
- I have deliberately waited awhile before reacting publicly to "the C++ MacApp
- thing," as George Bush might have called it. There have been a lot of very
- sharp insights shared since the news broke in Phoenix, so allow me to summarize
- what I see as points of general consensus, sprinkled with my own observations.
- Hopefully, this can help build a coherent strategy for how to proceed from
- here.
-
- • The whole issue was introduced in an unfortunate way. Granted, it was
- probably easier to plunge ahead with what was perceived as a necessary change
- than to deal with preemptive public critism, but even so the word "bombshell"
- is not inappropriate in describing the manner of introduction. This, after
- all, is a community that has remained loyal through inadequate staffing of the
- MacApp team, interminable delays in releases, numerous "betas" that change
- rather dramatically from one version to the next rather than simply fix bugs,
- inadequate documentation, and other hardships. We all know that MacApp is
- worth the wait and the effort. There is nothing to fear in advance discussion
- of even controversial subjects in this particular community. Instead, at a
- time when everyone should be formulating strategies to use the great features
- of MacApp 3.0, we have had rumors, intrigue, and rebellion.
-
- • MacApp 3.0 does, indeed, look like a leap forward. Lost in the language wars
- has been appropriate notice of the great strides made by the MacApp team.
-
- • The abrupt switch to C++ will, in fact, cause hardships in certain circles.
- Some are affected by the language switch, many by the perceived need to drop
- the Think environment. It is not yet clear the extent of this hardship and the
- good folks at Apple are hard at work gathering precisely that information.
- However, this again is information that should have been taken into account
- before the announcement, not gathered afterwards.
-
- • The ultimate change to C++ surprised no one that I know of, including diehard
- Pascal fans. We all recognize business realities, and if Macs are to become
- popular in corporate circles, thereby increasing the available market for all
- of us, C++ will have to be an integral part of the reason.
-
- • Yes, it is a laudable goal to not release source or ever worry about it. If
- object-oriented systems software is ever to become a commercially viable
- industry, companies that invent great software will need to protect it.
- Furthermore, having a class library which is usable without source indicates a
- high degree of discipline in its construction, quality documentation, and
- stability of the class hierarchy. Put another way, by building for a
- source-free environment, you create better software.
-
- • HOWEVER, no one knows today how to adequately document object-oriented
- systems to the extent needed for a source-free MacApp. Furthermore, no one
- really knows how to design stable class hierarchies without resorting to lots
- of real-world hammering, and the MacApp team, for all their tremendous
- achievements, has not established a track record of bug-free or almost bug-free
- software (yet). This is not a knock at the MacApp team; done properly, quality
- assurance on a 50,000+ line MacApp 2.0.1 is probably a comparable effort to QA
- on a 1,000,000+ line program in most other arenas. Object-oriented technology
- today is a curious mixture of ancient tradition and infancy.
-
- The upshot of all this can be summarized into these points:
-
- • The switch to C++ is inevitable. The questions are how and when, not if.
- • Regardless of language, MacApp 3.0 is going to be a great benefit to
- everyone.
- • Apple needs to do some damage control with those who will be impacted.
- Loyalty cuts both ways, and the MacApp community has been intensely loyal to
- Apple over the years. This could be done by maintaining 3.0 in Pascal, whether
- exclusively or not, or by speeding the development of Apple and third party
- Pascal products that will be C++ friendly.
- • Source code will continue to be needed until better documentation is
- available and code quality, measured by stability and lack of bugs, is
- increased.
- • In the future, Apple needs to keep to its tradition of open discussion.
-
- Regards,
- Jeff Alger
-
-