Talk:Transaction pattern (computer science)

From Wikipedia, the free encyclopedia

[edit] Cleanup needed

This article needs a bit of a working over to make it follow the manual of style. In particular, it needs:

  • A more concise leader (a couple of paragraphs vs. the current essay);
  • Editing to remove first-person discussion;
  • Better sectioning;
  • References.

The article name should probably be changed to "Transaction pattern" or "Transaction pattern (computer science)" to follow Wikipedia's naming conventions. --Allan McInnes (talk) 16:37, 16 May 2006 (UTC)

[edit] Original Research

This is entry is a summary of a self published work: http://www.tonymarston.net/php-mysql/design-patterns-are-dead.html

If this turns out not to be original research than I would recommend changing the title. I can't find any similar uses of term "Transaction Pattern" on the web. (Not counting Tony's work of course.) However, "Transaction Pattern" has been heavily coined as an alternate name for the command pattern. It is first referenced this way in "Patterns: Elements of Reusable Object-Oriented Software" by Gamma, E., Helm, R., Johnson, R. & Vlissides, J. (1995). Two additional references to this can be found at the following urls:

Mjlivelyjr 21:35, 2 June 2006 (UTC)


This thread at "Joel on Software" would seem to confirm that "transaction patterns" are original research on Tony's part. Specifically, I think that the following comments by Tony are telling (note that these comments are part of a thread discussing Tony's article Design Patterns are dead! Long live Transaction Patterns!):
  • "If what I have described as "transaction patterns" already exists under a different name, the what is that name? Where is it documented?"
  • "In my article I specifically identify the problem domain that I am addressing, so if your problem domain is different then my solution will not be appropriate."
  • "What I have described in my article in something which is not confined to common Lisp, but which could be applied to ANY computer language. Indeed, I have already applied these principles to two totally different languages."
--Allan McInnes (talk) 22:19, 2 June 2006 (UTC)


The fact that there is a design pattern with "transaction" in its name does not mean that it is exactly the same as my "transaction patterns". Design patterns exist at a low level and several may be joined together to form a working end-user transaction. Transaction patterns (note the plural, as there are different patterns for different circumstances) exist at a much higher level and may be comprised of any number of low-level design patterns, but from the user's perspective they are totally irrelevant. Each user transaction has a user interface, so the only things of interest to the user are (1) what does it look like? and (2) what does it do? The programmer, in his turn, takes these two points and turns them into a working program. If the application framework has transaction patterns built into it then it makes the programmer's job a great deal easier as he does not have to bother about which group of design patterns to use but which single transaction pattern to use.

Tony Marston 11:16, 19 June 2006 (UTC)

This is interesting, but does not address the point that your "transaction patterns" are original research, and thus don't belong on Wikipedia. --Allan McInnes (talk) 15:38, 19 June 2006 (UTC)


OK. I'll remove the entire entry.
Tony Marston 15:08, 23 June 2006 (UTC)
Retrieved from "http://en.wikipedia.org../../../t/r/a/Talk%7ETransaction_pattern_%28computer_science%29_8b6d.html"