[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [oiic-formation-discuss] Draft Interoperability and Conformance TC formation proposal (0.2)
On Sat, Jul 26, 2008 at 6:39 PM, Gordon MacGinitie <gmacginitie@gmail.com> wrote: > To all > > If this small group were to do certification or any testing or provision of > testing software it would be faced with never ending accusations of > prejudice, discrimination, bugs hidden in the software and on and on. There > is to be only one OIIC TC but there will be many people that offer actual > tests or testing software; the specification of that which is to be tested > can be very open and free of bias and competition among the testers will > quickly bring any biases to light. > > A small group that meets by email would be appropriate to the job of > (slowly) creating guidelines and specification and not so to writing > software or running a testing business. > Issue is validation of reports. Many bodies with different quality tests you cannot validate reports without the tests used. People will say old tests were used or so on. Any report generated need to be linked to a exact copy of the tests that generated the data used in it. These could be applications. This way any dispute can be confirmed as true or false and got over and done with. Reason upset implementer can either disprove the test, except the error in program or confirm a documentation error. If its left to wild west. One implementer will say I passed X Y Z test. Another one will say passed A B C test covering same section yet report writter used D E F test and called us wrong. So now we have a test compatibility problem. Issue now is that D E F updated there test since the report writer got them and thiers is now defective. Even that in the first place it was right. Now the reporter becomes a laughing stock and due to TC rules cannot defend self. If the TC is going to write reports on conformance to ODF there is no option bar to store the tests and be able to ship them when needed. Even in some cases send out acid test groups of tests for broader testing to make sure its not still existing because of updates and the like. Basically storing and sending out tests is a requirement to be above question. Being able to code tests more like build test kits from existing tests is also key to the process at times. It is too soon to be setting these rules in too hard of stone. Flex has to be left. Peter Dolding
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]