This author has a blind spot
|
| Author | Content |
|---|---|
| tadelste Sep 12, 2005 8:53 AM EDT |
Open source's weakest leak is testing according to the writer of this article. I've been working with open source and free software projects for less than a decade, so I may not be up to speed on this subject. But, testing in my experience has been one of the key strengths along with innovation and the sheer mass of development. For someone who has only done proprietary software development - I can imagine that they might mistake open source testing for something else - especially if they're used to hiring QA firms who do nothing but test. His assertion seems uniformed to me. |
| dinotrac Sep 12, 2005 9:25 AM EDT |
I don't know, Tom. Obviously, SpikeSource has a product/service to sell and I don't begrudge them that. As with any product sales, a little bit of puffery is to be expected. At the same time, I'm not at all sure that I share your rosy view of open source testing. Now, before I say anything, we need to remember that testing of proprietary software is all over the map. Some companies take a much more serious view of quality than others. That said, in the bazaar world, lots of people may be working on code that may or may not create side effects for other people working on related code. Platform ports may be done by limited groups of users for specific purposes and not tested beyond that set of needs, etc. You know it's true, Tom. It's not so bad if you run Linux on an *86. I've found problems running Linux on X86_64, and considerably more problems running AIX on power. Some projects are very scrupulous about regression-testing and the like, some are less so, and all tend to get a little less scrupulous as they move on to "other platforms". By and large, the sheer number of users + code availability tends to get problems addressed. Still, consider all the efforts by the kernel team to get source-code control handled properly, the KDE team likewise, etc. It's hardly beyond the realm of reason to suggest that testing (especially boring dumb old regression testing) might benefit from some similar love. |
| tadelste Sep 12, 2005 10:35 AM EDT |
Dinotrac: Rather than look at OSS development as a perfect model, I'm thinking strictly in terms of proprietary versus open development; waterfall versus iterative. Every product and project manager will have his or her own ideas about testing. In my world, I furnish specs and change orders to testors and rotate them. I'm not suggesting that we don't test - I just think it isn't necessary an inherent weakness since I believe testing gets done. |
| dinotrac Sep 12, 2005 11:10 AM EDT |
Tom: I agree it's not an inherent weakness. The logical fallacy that makes the article less than honest (well, it IS a sales piece) is the presumption that proprietary testing is any better, or, for that matter, that no open projects are able to do scrupulous far-reaching testing. Try selling that last proposition to the postgresql developers!! Still, it is interesting to wonder if there's a reasonable way to apply some of the bazaar development lessons to testing. |
| tadelste Sep 12, 2005 1:00 PM EDT |
Agreed! |
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!