Slashdot Mirror
OSS Developers Provide A Glimmer of Hope
sebFlyte writes "In a recent speech at the ACCU conference in Oxford, software design guru James Coplein said that unless consumers start demanding more and putting up with less crap from software firms, the quality of proprietary software would keep spiralling down. He was full of praise for open source though, saying 'The complementary, independent, selfless acts of thousands of individuals can address system problems -- there are thousands of people making the system stronger.'"
... that the general quality of EVERYTHING is on a downward spiral. Relentless commoditisation is forcing everyone to work on lower margins and wider tolerances.
The statement only seems partly true for those who are really, really big and have a kind of monopoly (you know who I mean). For smaller, niche software, they have to make it good and even better for each version or the customers will demand something better by buying something else. So a company who neglects their customers needs will go broke and disappear, giving those companies who listen to their customers a higher market share. And even the biggest software companies are seeing that their market share drops when they do something bad, because even the biggest monopolies face competition from things like Linux.
9/11: Never forget it was a false-flag operation
The problem is that users don't give a flying fuck about licenses. Mainly due to the reason that they're written in terse legalese, and they don't affect the end-user. Licenses are seldom enforced on most end-users, and as such, nobody cares if they're entering into a contract in exchange for their first-born.
Additionally, the users don't have too much choice (referring to a certain Office suite and operating system) - installed base can be a good thing for the company owning said software. The ones who do make a choice tend to be more educated about software licensing, but frankly, as long as the end-user is not affected by software licensing, it's like downloading music - nothing's going to change until the end-users are attacked.
I've been reading today about Iterative and Increment Design (IID) which is based around the principle of breaking a major project up into smaller iterations (of say 1-6 weeks) and at the end of each of these, integrating all the code and demonstrating it to the customer, whose feedback is used to adapt the product development in order to eventually end up with a final release which is useful.
It can even be taken as far as evolutionary delivery, which requires that the software be released into the market, and the feedback from that used to decide what will be in the next release. The time scales of this are much shorter than say, Apple releasing Panther and then Tiger, so not to be mixed up with major product releases.
I only wonder whether the success of Linux as a household brand is compromised by the fact that non-proprietory software does not follow IID and hence doesn't actually deliver what is the customer wants, but in fact what the developers think the customer wants. I know that Microsoft are very much for beta testing on thousands of individuals which is a step closer to this, but from the serious delays in Longhorn, it's also true that maybe they have missed the point as well.
There's no doubt the functionality is there in Linux as the guy mentions but I'm not so sure that the customer really fits into Linux like is required when moving beyond the waterfall model...
Was having a conversation with an "I only do Windows" manager who was trying to be agreeable with me (knowing I'm an "I don't do Windows" contractor). He said he thought Microsoft would be helping to boost Linux growth by their recent push to enforce licensing.
As much as I appreciated his sentiments, I had to respectfully disagree. I illustrated my perspective by pointing out how we had both spent the last six hours cleaning off spyware from the reception desk PC of one of his client's. (He needed my knowledge of Knoppix to pull important documents off the workstation, just in case.)
To summarize, I said, "People will put up with incredible amounts of discomfort and expense, rather than learn something new." I think Microsoft has figured this out, long ago. I'll add that it doesn't help that most business software (e.g. Quicken, QuickBooks, Point, etc.) is built for Windows and that that fact will probably never change.
Linux in the embedded world will grow. Linux in the server world will grow. Linux for the business desktops won't. Not for a long while -- if ever. After watching my friend scrape spyware dung off the Windows' registry, for hours, oh, how I wish it were not true.
The underlying problem is that there is a short-term (and perhaps long-term) commercial advantage to shipping buggy, poor quality software "today" rather than higher quality software "tomorrow".
OSS has no advantage to shipping software before it's ready-- This can sometimes backfire, because if the OSS developers stop making updates/bugfixes, either other people pick it up, or the project is stalled. A commercial company would still need to do at least major bug fixes if they want to keep customers coming back for version 2.0.
Also, some projects just don't work well with the OSS model. Games, for instance-- some of them are more like movies, and needs $$$ to back them.
Wer mit Ungeheuern kämpft, mag zusehn, dass er nicht dabei zum Ungeheuer wird. --Nietzsche