Talk:Termination analysis
From Wikipedia, the free encyclopedia
[edit] Only valid answers?
The assertion that the only valid answers are "definitely terminates" or "don't know" is patently false and a misinterpretation of the Halting Problem. Consider a simple program:
010 GOTO 010
Trivially, this program will never terminate. What must be true is that any given method must have at least one program that it cannot determine, and so must any union of finitely many methods. Robert A.West (Talk) 00:00, 1 September 2006 (UTC)
This above^ is why we need some knowledgeable folks to fix this "article". I pulled this definition down today (1 Sept 2006) from the FOLDOC and just pasted it in to get something going.
I was bothered by the following statement too:
- "For example, evaluation of a constant is bound to terminate."
- I don't quite know what "evaluation of" means.
- If "evaluation of" means "to determine" as in "to create" a constant such as pi I don't see why the calculation creating pi has to necessarily terminate
- And even if "the method" starts to calculate pi (apparently) successfully, will it erroneously terminate, say after calculating only 1,258,433,705 digits? In other words, do we have an "ineffective" algorithm because of premature termination?
Am I missing something obvious here? wvbaileyWvbailey 02:47, 1 September 2006 (UTC)
If someone wants to "unstub" this article feel free to contact me. I am part of the AProVE project (http://aprove.informatik.rwth-aachen.de) and might be able to explain details and important facts. cotto@i2.informatik.rwth-aachen.de —Preceding unsigned comment added by 137.226.108.169 (talk) 22:50, 29 March 2008 (UTC)
