On Wed, Mar 26, 2008 at 12:44 PM, Ovid <[EMAIL PROTECTED]> wrote: > --- Gabor Szabo <[EMAIL PROTECTED]> wrote: > > > I wonder if it would be possible to take the existing .*Unit > > libraries > > of Java and .Net and > > create some wrapper around them (or a replacement) so people with > > existing tests > > written in those testing system would start producing TAP results. > > In theory, this shouldn't be too hard. In practice, there are issues. > I'm going to write a bit of negative stuff here, but that's not to say > that we can't or shouldn't do this because I think this is a great > idea. It's just that the testing worlds involved have different views > on how things work and as we all know, these views are often > "religious" in nature.
Just as the choice of language is religious in many cases. I have a feeling for example that the Python community will have a very hard time to accept Parrot even if it has advantages over other things they might have. Just because it comes from Perl people. The PHP community seems has less negative feeling towards Perl. I think this is also reflected in the adoption of TAP. It seems PHP people are more ready to use TAP than Python people. .NET and Java people might have no such negative feelings towards Perl. They just think it is not strong enough compared to their enterprise language. That means getting them accepting TAP is quite an uphill battle. Making it a drop-in replacement or a wrapper around their current system might be a key issue in adoption. The other one would be a nice GUI for TAP aggregation and reporting. I have copied some of the responses to the "why use Perl for testing?" question to http://perl-qa.hexten.net/wiki/index.php/Why_use_Perl_for_Testing It would be nice if others would help putting together a good document. regards Gabor
