Ordering, Was Pair! thread

Hi all, as I (and Joel) see it, RT decided to support a default ordering in Rebol. That ordering is used for sorting purposes. Properties of the default ordering: 1) some values aren't ordered: a: [] insert/only a a 0 b: [] insert/only b b 0 sort reduce [a b] 0 2) The ordering looks as being non-sharp and weakly antisymmetric with respect to the Equal? function (but not W.R.T. the Same? function). 3) The default ordering orders pairs

Hi ladislav, Maybe I / we have as Joel termed it a too puritanical view with regards too such things as language consistency or notions of "correctness" with regards to mathematical or numeric behaviour & numbering systems. Gregg talked about "conceptual integrity" and I agree that conceptual integrity is very important. RT often like to dismiss the highlighted inconsistencies and confusing or missing REBOL behaviour or features as "too confusing for newbies" or "only important in the rareified atmosphere of academic theories" as a justification for the current REBOL behaviour, whether this is to buy them time to bug fix or redesign or simply ignore as they do not consider the matter imprtant then only RT know the answers to that. REBOL currently have lots of areas that I consider to be lacking in "conceptual integrity". I suppose this puts me in the position of being "noise" if I repeatedly raise these concerns here, so I consciously try not to, maybe Iam too confrontational at times so I've been trying to say less on this list and don't wish to offend or upset people, but sometimes I just get frustrated when I see what REBOL could be and the reasonings given for why it is how it currently is. Please remember it is only because I DO actually care about REBOL as a technology. I also care about REBOL Technology and hope they succeed but as I see it there is a mismatch between their core skillset which is developing the foundation REBOL technologies and what I percieve is the market need for a REBOL "SOLUTIONS" company to provide and maintain programs and applications built on core REBOL technology which I believe should be an "open" standard and multi-vendor and preferably open source to enable the language / technology to fully grow and provide the functionality and explore areas which RT Inc. currently do not have the resources or interest in pursuing. Python is a perfect role model for REBOL in my opinion. Python has wonderful diversity and a whole array of modules and extensions. REBOL is a better language in my opinion but it's growth and adoption is currently limited by the RT Inc. bottleneck. That's my two cents worth. I accepted a long time ago that REBOL is not going to be open sourced but it doesn't make it any less frustrating when I see the rate of progress that other language technologies are making in comparison with the slower pace and apperently frequent directional changes that RT make with regards to product development / strategy. I see similarites between what RT Inc are trying to achieve and what BEos Inc. tried to do and failed. that's my two cents worth. Mark dickson Thats all. Mark Dickson In a message dated Fri, 8 Feb 2002 6:35:40 AM Eastern Standard Time, "Ladislav Mecir" <[lmecir--mbox--vol--cz]> writes: