On Sat, Apr 17, 2010 at 3:49 AM, Corey <[email protected]> wrote: > On Friday 16 April 2010 21:29:44 [email protected] wrote: >> > Messy, with high levels of noise-to-signal - certainly... but absolutely, >> > astoundingly productive and in constant motion. >> >> In my opinion, most of the output from the Posix developers is trash. >> It's the equivalent of a cancer, polluting the body with poisons. >> Somewhere in the mix there will certainly be something of value, but >> it is well hidden by the bulk of the production. The few jewels are >> also corrupted by the manner in which they need to be delivered, >> namely the autoconf stuff. >> > > Understood. Though I don't share your opinion quite to the degree that > you expressed. Additionally, I have no desire to debate subjective > perspectives of the overall net usefulness of POSIX, let alone autoconf - > everyone, of course, has their opinions and experiences, favorable or > otherwise. > >> If you consider things more objectively you will also acknowledge that >> very little new is being created, but rather many old things are being >> "improved" upon (regurgitated) in manners that consume more and more >> computing cycles and deliver less and less performance. >> > > Again, I'd prefer to not to debate the ratio of good software vs. trashy > software, or to debate what's new and useful vs. merely regurgitated and > worsened. Though it's certainly a perfectly interesting topic. > >> Consider further the following: porting GCC/G++ to a new platform >> rather than Linux is almost inconceivable, porting more and more Linux >> software to a compiler suite other than GCC/G++ is equally >> inconceivable. If you can't see anything wrong with GCC's bloat, the >> dead end it leads to, there is little reason to argue with you. >> > > Finally, regarding this mention of gcc - the "Plan X" in my mind's eye > would far prefer LLVM/Clang to gcc, for precisely the reasons you > point out. (I've been considering the prospect of implementing a > kencc dialect for the clang c front-end). > > (I'm using "Plan X" in the sense I mentioned in the original post - i.e. > I'm _not_ suggesting that the official releases of Plan 9 proper should > introduce the platform changes under discussion. "Plan X" means: > any alternative expression of the Plan 9 operating system. Also, I'm > using the phrase "my mind's eye", in order to stress that this is all just > speculative, science-fiction) > > Regarding the POSIX situation - a "Plan X" of my mind's eye is not > concerned with fighting that particular battle. > > The basic wild-eyed premise, is that an alternative Plan 9 distribution > which "features" a native, more "POSIXy" approximation than APE, in > addition to a native compiler that supported a larger number of languages > and C dialects than 9c - would lead to a much more broadly comfortable > environment for a greater number of general developers and users. > > The theory, is that the Plan 9 implementations of the following concepts: > > * 9P > * mutable namespaces > * union directories > * ubiquitous fileservers > * transparent distributed services > * etc > > ... are simply - by far - much more important and practical to a greater > number of people than these other prominent Plan 9 idioms: > > * radical frugal simplicity throughout the entire system > * a stance against POSIX and other standards > * a stance against alternate programming language paradigms > * a strong bias towards a particular form of user interaction with the > system (i.e. acme, rio, etc) > > > Peace > > > >
still too long.
