"Keep your coding high and your expectations low." - I like, I will add
that to my bucket of nuggets.
I wrote the output formatting in a modular way so that more formats can be
added in the future.
The one output format compatible with other testing frameworks that I have
already added is TAP,
I actually had not thought about a migration strategy for tests from
different frameworks. Hopefully
this would be something that can be done with a community effort.
I have emailed the mailing lists of the developers of most frameworks
mentioned in Phil's blog and
a few others. I guess now we wait and see.
Thanks a lot for the message of encouragement and support. Likewise all the
best to you with your
library. I actually haven't heard of it so maybe you can post a link on
this mailinglist just incase its
something I might need down the line.
On 10 November 2012 20:09, Faré <fahree@...> wrote:
> Dear Tapiwa,
> > It was at the point I stumbled on François-René blog, I decided to take
> > the challenge and consolidate the unit testing frameworks. I have no
> > experience in working on community efforts so I came here because I
> > this is were I can get the most useful advice and constructive criticism.
> Thanks a lot for undertaking this project!
> Keep your coding high and your expectations low.
> > I have written a framework that aims to consolidate all the major
> > of the frameworks mentioned in Phil's blog. You can find it on Github
> > https://github.com/tgutu/clunit
> I haven't personally investigated the space of testing frameworks enough
> to make any claim as to the appropriate thing to do,
> whether it is creating a new framework indeed, or
> starting from one of the existing ones.
> (Hell, in my own pet domain, data structures,
> I have indeed started a new framework, LIL, so I won't cast any stone).
> However, let me just say that unless you can indeed convince
> a sizable number of users and developers that
> your new framework is indeed the Right Thing™,
> for which you will need to have very good arguments,
> then starting a new framework will have achieved the opposite of the goal,
> i.e. more fragmentation rather than more consolidation.
> Also, not mentioned on Phil's blog are some frameworks that can produce
> an output compatible with testing frameworks for other languages,
> which allows for unified interfaces to test and test statistics.
> I haven't used any of them, just heard about them.
> > I wrote a blog on the development of the framework and reasons for it
> > http://ml.sun.ac.za/2012/11/09/developing-a-unit-test-framework-part-1/for
> > those who are interested.
> Whatever happens, please don't get discouraged.
> If you do one library well and it indeed becomes THE library for the topic,
> you will have done more than if you do ten bad libraries.
> Nevertheless, a big part of the challenge of a community is
> to get the conversation going with users and with developers of other
> I suggest you contact the developers and/or users of each existing
> and get familiar with how they see things.
> Finally, it is interesting to think of an eventual transition strategy.
> Is it a matter of rewriting clients to use your new interface?
> Is it a matter of writing compatibility systems that express the old
> in terms of the new one?
> Best of luck!
> —♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics•
> Economics is the plural of Morality.