Talk:Data-flow analysis
From Wikipedia, the free encyclopedia
This should probably be moved to wikibooks, as it is not an encyclopedic article. Thue | talk 15:07, 15 Jul 2004 (UTC)
There should be a definition of "dataflow analysis".
We can say something like : "Data flow analysis is a process for collecting run-time information about data in programs without actually executing them" --Cocothebo
Technically, it's data-flow (with a hyphen) analysis. FAdmMatt
- You seem very certain of this. Why is "control flow analysis" always spelled without the hyphen? Certainly, the Dragon book uses the hyphen, but I find that most papers on the subject do not (some use a space, some neither a space nor a hyphen). I'd really like to know the definitive answer. --Mike Van Emmerik 12:09, 14 February 2006 (UTC)
[edit] not enyclopedic
Agreed that this is not encyclopedic, but disagree that it should be moved to wikibooks. If I knew anything about data-flow analysis, I would write it here, but I don't (which is why I am here). Suggested sub-topics:
- what it is
- why it is used
- history/background
the FolDoc entry might be a good place to start. --Lenehey 22:36, July 28, 2005 (UTC)
[edit] Reverse postorder is not the same as preorder!
It's really scary that this article has claimed since its creation in 2004 that reverse postorder is just another name for preorder. So much for the many-eyes philosophy. This is the worst kind of mistake an article can have, because once you get an idea like this into people's heads it tends to be very difficult to get it out. I fixed the problem, but I'm placing an erratum here in the hope of catching the eye of a few of the people who were misled by the article. -- BenRG 15:04, 2 September 2006 (UTC)
- Hmm, actually on rereading that paragraph I'm not sure what it's talking about. What does "except when the successor is reached by a back edge" mean? Is it just talking about the fact that you don't revisit nodes you've already visited? Unless there are objections, I think I'll rewrite this paragraph to simply say that reverse postorder is the reverse of postorder. -- BenRG 15:15, 2 September 2006 (UTC)