puzzle
in [General Programming]
|
Prev: Algorithm book recommendation?
Next: TaskScheduler.dll
From: Peter Ammon on 20 Jun 2005 05:40 Randy Howard wrote: [...] > > > I think it's the same compiler that mystically hoists strlen() > calls out of the top of for loops. It's never been demonstrated, > but he swears that "any decent compiler" will do it. Apparently, > there are no decent compilers in nilgewater land. Not to rise to the defense of Nilge here, but gcc 4.0 at least will hoist strlen() like that. -Peter -- Pull out a splinter to reply.
From: pete on 20 Jun 2005 05:48 Christopher Barber wrote: > > spinoza1111(a)yahoo.com wrote: > > > > On the one hand, you can introduce constraints. > > On the other, you can find a deus ex machina. > > Therefore the theory does not categorise > > algorithms once and for all, it allows us to speak about > > properties of algorithms. > > > > But owing to corporate surveillance and its Benthamite potential, > > the speech is silenced by posters who have to treat the > > theory as a set of dead answers. > > You keep spewing this "corporate surveillance and Benthamite > potential" mantra. What does it mean? There are five distinct varieties of Benthamite (green, red, gold, blue, and white), the first three of which are toxic to Superman. -- pete
From: CBFalconer on 20 Jun 2005 11:21 spinoza1111(a)yahoo.com wrote: > .... snip, leaving the following foul smelling lump ... > > I did hypothesize as to whether an optimizer could do in some > advanced way algorithm replacement but this was clearly in the > context of an understanding that it cannot be done. By doing a little digging I found the following, which is quoted verbatim. >> is Barber, > is Nilges: >> Why bother with optimizing a bad algorithm? The only time you >> would use such an algorithm is when n is very small so there is >> no need to perform any other optimizing tricks. When n is >> large you had better be using an O(n) algorithm. > > This is what optimizing compilers do, mate. And in software > maintenance you need to know how to make low-risk code > improvements to "bad", Commie, terrorist, child molestor > algorithms. Clearly showing the quality of the Nilgean denial. Rhymes with Augean, I wonder if that is a coincidence? Either requires much shoveling. Just for clarity, my quotes are normally in complete paragraphs. This tends to better define context than simply quoting one or two words. However this business closely resembles the training of puppy-dogs. Rubbing their noses in it doesn't usually improve their behaviour - they just don't understand. So I pose the question to the readership: Are there any such redeeming qualities that attempts should be made to train the Eddie entity? Or should we enjoy the relief that comes from ceasing the banging of foreheads on brick walls. -- Some informative links: news:news.announce.newusers http://www.geocities.com/nnqweb/ http://www.catb.org/~esr/faqs/smart-questions.html http://www.caliburn.nl/topposting.html http://www.netmeister.org/news/learn2quote.html
From: Christopher Barber on 20 Jun 2005 12:25 spinoza1111(a)yahoo.com wrote: >>Indeed. Out of curiousity, Tisdale, how would endianness, or any >>ordering, affect any Boolean operator? > > What part of shift don't you understand, Homer? Bit shift operators are not affected by the endianess of integer representations.
From: Christopher Barber on 20 Jun 2005 12:28
CBFalconer wrote: > So I pose the question to the readership: Are there any such > redeeming qualities that attempts should be made to train the Eddie > entity? Or should we enjoy the relief that comes from ceasing the > banging of foreheads on brick walls. I find his posts amusing, although presumably not in the way they were intended. However, I don't think that he is "trainable" and I suspect that anyone that expects him to change is bound to be disappointed. However, we cannot simply let him make false claims without at least some refutation lest someone new to this group actually believe them. - Christopher |