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


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index]
Re: [xml-dev] Testing XML don't use xUnit

On Wed, Apr 10, 2013 at 6:22 PM, Fraser Goffin <goffinf@gmail.com> wrote:
>> But XMLUnit doesn't let you do that because it requires an absolute path
> What makes you think that ? What if your XPath starts with
> //MynodeUnderTest. Is that still a problem ?

Permit me to summarise and withdraw from  being drawn in to a
perpetual response cycle.

The idea of not exposing navigation paths in test code to the QA
community is not new. They write books about it, give names to
techniques to avoid doing it and stage training courses and write
conference papers about it. Of course they also make money from it.

This was common knowledge long before unit testing frameworks and
automated unit testing became de rigeur in the development community.
I personally discovered it on a job 21 years ago.

As part of my degree I read alot of research about evolutionary
computational methods for automated test case generation. I was
shocked to find that the research community was using 100% code
coverage as a benchmark for evaluating the efficacy of their
algorithms. Shocked because I knew  that that was a very weak
criteria. That is if your algorithm could automatically generate test
cases that gave 100% code coverage it really hadn't achieved much (to
a large extent the same thing is a by product of test driven
development) so such research was never going to have much industrial
relevance. That much was obvious to me because I had a background as a
QA dev going into university.

As a QA dev that came from the development community I noticed things
that seemed weird by virtue of my prior experience. Why are you paying
5k for a box of tricks that has fewer capabilities than Visual Studio
which is a 10th of the price. If you want to run a test against the
underlying database why don't you just write SQL?

So when you have oscillated between the various communities (academic,
dev and testing) you notice situations that have arisen from a  lack
of cross fertilisation of ideas between the respective communities and
you notice that instinctively. I would emphasise that it applies
equally to all three strands I have been involved in. They all
reinvent wheels and don't take advantage of what is common knowledge
beyond the confines of the community that they inhabit.

So I've tried to share an instance of this in this thread. I used to
have  similar arguments with QA people - why do you tolerate products
like WinRunner and QARun that come with proprietary scripting
languages why don't you demand products that hook into mainstream
programming languages - and they used to come up with a whole set of
justifications for justifying the status quo - not dissimilar to the
genus of some of the arguments I've seen here.

So I'll just sign off with this.  That it is always worth looking
outside to see how other communities have dealt with problems similar
to the one you are contemplating.

Hopefully we can all agree on that.


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

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 1993-2007 XML.org. This site is hosted by OASIS