Re: Test suite output format

On 5 September 2012 17:06, Yves Savourel <ysavourel@enlaso.com> wrote:

> Thanks for working on this Leroy,****
>
> ** **
>
> A few notes:****
>
> ** **
>
> **-   **It seems there is still a XML header at the top.****
>
> **
>
Sorry that was a mistake with my output i will get rid of that.

> -   **It may be easier to open the files if the extension is .txt
>
I will make the files .TXT not .XML

> ****
>
> **-   **I think we said we would not output outputType for that format.
>
Didn't know about that but now that i looked at your outputs i see that . I
will remove that now.

> ****
>
> **-   **We may want to quote the resulting values (although if it’s too
> complicated to mix attribute syntax and tab-delimited any single-line
> syntax would work, as long as we agree upon one.
>
It should be simple enough to do whichever you think is best.


Thanks,
Leroy

> ****
>
> ** **
>
> Cheers,****
>
> -yves****
>
> ** **
>
> ** **
>
> *From:* Leroy Finn [mailto:finnle@tcd.ie]
> *Sent:* Wednesday, September 05, 2012 9:52 AM
> *To:* Felix Sasaki; Multilingual Web LT-TESTS Public
> *Cc:* Dave Lewis; Yves Savourel; Dominic Jones
> *Subject:* Re: Test suite output format****
>
> ** **
>
> Here is a small sample of the tab delimited output for translate.****
>
> ** **
>
> Leroy****
>
> On 5 September 2012 13:02, Felix Sasaki <fsasaki@w3.org> wrote:****
>
> Great, thanks a lot, Leroy.****
>
> ** **
>
> Best,****
>
> ** **
>
> Felix****
>
> ** **
>
> 2012/9/5 Leroy Finn <finnle@tcd.ie>****
>
> Hey Felix,
>
> I have the new text output ready to go I have basically put the output
> through another XSLT page to make it produce text output as described
> by Yves. I will add in details to the page about the textual output
> when i am adding in the new output files to the webpage (while also
> updating all the zip files).
>
> Thanks,
> Leroy****
>
>
> On 4 September 2012 11:02, Felix Sasaki <fsasaki@w3.org> wrote:
> >
> >
> > 2012/9/4 Leroy Finn <finnle@tcd.ie>
> >>
> >> On 3 September 2012 22:09, Felix Sasaki <fsasaki@w3.org> wrote:
> >> > Thanks for the nice feedback, Dave.
> >> >
> >> > One question I have is how to get feedback from the implementors, at
> >> > least
> >> > the people who have to do implementations in LT-Web. I assume that
> >> > except
> >> > Yves and I nobody has looked into this discussion. How can we assure
> >> > that
> >> > they will be able to deal with the test suite?
> >> >
> >> > Another question is the hosting and the generation of test files and
> >> > output
> >> > files. Would it be OK to host the test suite on the w3c site? For this
> >> > site
> >> > there is some guarantee that it will be available in a few years. Can
> >> > TCD
> >> > guarentee that for
> >> > http://phaedrus.scss.tcd.ie/, or should we move the tests?
> >>
> >> We can host it here for the time being but it maybe better to host it
> >> on the W3C especially when its finished where it can be guaranteed to
> >> stay up. So if you could organise hosting that would be great.
> >>
> >> > If yes, is there a machinery for generating the tests, or is both
> input
> >> > and output
> >> > handwritten?
> >>
> >> The input test file is handwritten and gotten from the draft examples.
> >> The output files on the other hand are generated from the test suite
> >> code that you sent me previously it has been updated though to include
> >> the ITS 2.0 data categories.
> >
> >
> > Thanks for the info, Leroy. I think we don't have to put the machinery
> for
> > generating output files on the w3c server, but once the data categories
> are
> > stable we may want to put the test suite itself (including the new
> output)
> > to the w3c server. Do you think you could document the "new" (text based)
> > output format and have it generated by the September f2f? That would be
> > sufficient.
> >
> > Thanks,
> >
> > Felix
> >
> >>
> >>
> >> Thanks,
> >> Leroy
> >>
> >>
> >> > Best,
> >> >
> >> > Felix
> >> >
> >> >
> >> > 2012/9/3 Dave Lewis <dave.lewis@cs.tcd.ie>
> >> >>
> >> >> Hi Felix,
> >> >> that's great, as we also need Leroy to help implement some of those
> >> >> multi-data category roundtrip scenarios I mentioned earlier over the
> >> >> next
> >> >> few weeks.
> >> >>
> >> >> We really appreciate all you help in checking the test suite as it
> >> >> develops.
> >> >>
> >> >> cheers,
> >> >> Dave
> >> >>
> >> >>
> >> >>
> >> >> On 03/09/2012 13:57, Felix Sasaki wrote:
> >> >>
> >> >> Thanks, Leroy. No need to hurry, what we have now is sufficient to
> >> >> discuss
> >> >> the test suite format w3c internally.
> >> >>
> >> >> Best,
> >> >>
> >> >> Felix
> >> >>
> >> >> 2012/9/3 Leroy Finn <finnle@tcd.ie>
> >> >>>
> >> >>> Hey Yves,Felix,
> >> >>>
> >> >>> Thanks for your examples about this tab-delimited output.  This will
> >> >>> take
> >> >>> a few days to update the actually test suite and then to run all the
> >> >>> test
> >> >>> files so i don't think it is possible to get this all done by
> >> >>> tomorrow.  But
> >> >>> i can start working on it today and we could possibly aim for some
> >> >>> stage
> >> >>> next week to publish this output. But i will add in the output
> >> >>> description
> >> >>> on the ITS test suite page today based on this tab delimited output.
> >> >>>
> >> >>> Thanks,
> >> >>> Leroy
> >> >>>
> >> >>>
> >> >>> On 1 September 2012 13:59, Yves Savourel <ysavourel@enlaso.com>
> wrote:
> >> >>>>
> >> >>>> Hi Felix, all,
> >> >>>>
> >> >>>> > One question: for some data categories we would have a
> >> >>>> > lot of values for one node, e.g. loc quality issue in the
> >> >>>> > standoff version. How would the output look like
> >> >>>>
> >> >>>> Maybe simply the list of the expected values, in a pre-defined
> order,
> >> >>>> with an empty spot if the value is not present? See attached
> example.
> >> >>>> (I
> >> >>>> have not verified the actual results, but the format should
> >> >>>> illustrate the
> >> >>>> idea)
> >> >>>>
> >> >>>>
> >> >>>> > Also, some data categories have XML fragments (elements with
> >> >>>> > potentially attributes and (mixed) content inside). How to
> >> >>>> > represent these, e.g. what to do about whitespace stripping
> >> >>>> > for the XML fragments? I have no preference / idea, just asking.
> >> >>>>
> >> >>>> We should probably escape the line breaks to "\n" (and any tab to
> >> >>>> "\t")
> >> >>>> and escape any XML constructs. For example if a localization note
> is:
> >> >>>>
> >> >>>> <elem>text  with<tag/>
> >> >>>> And line-breaks</elem>
> >> >>>>
> >> >>>> We would output:
> >> >>>>
> >> >>>> its:locNote="text  with&lt;tag/>\nAnd line-breaks"
> >> >>>>
> >> >>>> cheers,
> >> >>>> -yves
> >> >>>>
> >> >>>
> >> >>
> >> >>
> >> >>
> >> >> --
> >> >> Felix Sasaki
> >> >> DFKI / W3C Fellow
> >> >>
> >> >>
> >> >
> >> >
> >> >
> >> > --
> >> > Felix Sasaki
> >> > DFKI / W3C Fellow
> >> >
> >
> >
> >
> >
> > --
> > Felix Sasaki
> > DFKI / W3C Fellow
> >****
>
>
>
> ****
>
> ** **
>
> --
> Felix Sasaki****
>
> DFKI / W3C Fellow****
>
> ** **
>
> ** **
>