When we produce XML, what degree of equivalence to the sample is required? According to which normalization rules should the equivalence be performed?
I ask because none of the language samples that describe what output they actually produce generate the exact text in the task description. —Dkf 08:19, 21 May 2009 (UTC)
Shouldn't the task be extended such that the document contains at least one attribute? -- Wmeyer 00:39, 13 January 2010 (UTC)
- Yes in principle, but it's a big round of breaking of existing examples to do. Is it that important? –Donal Fellows 10:32, 13 January 2010 (UTC)
- For scenarios like that, it makes more sense to create the "extended" version as a new task, deprecate this one, and forward-port as many examples as possible. (Do we have a deprecated task category? We're likely going to need one. ISTR creating a dep task template.) --Michael Mol 20:03, 13 January 2010 (UTC)
- I'm now convinced that it is probably not important enough to break examples or create a new task. Especially since XML Creation is a similar task that already deals with attributes. -- Wmeyer 20:17, 13 January 2010 (UTC)