OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

office message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re: [office] Formula: test cases


Patrick Durusau:
> PS: Would standards by mathematical associations count? Not ISO but 
> certainly similar in character.

I'm not claiming that many other organizations include test cases in their official specs. Pointing to "what others do" doesn't answer the question "what is the best approach."

I'll note that essentially ALL spreadsheet help systems do INCLUDE examples, which act just as our test cases - they clarify the meaning of the supposedly unambiguous description.  Microsoft's XML format includes examples (test cases) as well, though not in an automatically processable format, and I'm not sure if they are normative.

Also, the strictly mathematical functions (SIN, etc.) were generally the easy ones; there are LOTS of sources for high-quality, unambiguous definitions. The nasty ones are functions like the financial and date functions, which are easily described in ways that APPEAR rigorous but in fact are not.  Mathematicians are very much at home with rigor; in contrast, most financial users only know that they push the buttons and magic comes out. Since they're the users of financial functions, the definitions of many such functions turn out to be dreadfully ambiguous in the literature.

It's a question of tactics: Does a general definition ALONE guarantee that I have unambiguous text? Or does a general definition, combined with test cases, do a better job?  I am arguing the latter.

--- David A. Wheeler


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]