[ Date Index ][
Thread Index ]
[ <= Previous by date /
thread ]
[ Next by date /
thread => ]
On Thursday 08 Jan 2004 2:51 pm, Adrian Midgley wrote:
On Thursday 08 January 2004 13:18, Keith Abraham wrote:I know it's only "Microsoft Jack" but I think the following deserves
Not at all, the article is not as biased as other flame articles. Make sure we are clear on Free Software and Open Source before going further.
a knowledgeable reply. http://www.guardian.co.uk/online/story/0,3605,1117835,00.htmlDraft for my contribution to a reply.
My additions/amendments: Usability testing: Testing that a piece of software is portable to other environments, stable on other systems and understandable in usage. When a piece of Free Software is published, it is made available to a far wider range of environments, systems and users than any proprietary product and it is released earlier. Free Software is commonly released in beta if not alpha code condition because these usability problems are easier to solve when the program structure is still fluid. Proprietary software has limited in-house testing and is then thrust into variable environments as a fait-a-complit. Bugs are harder to solve because too many other functions, features and add-ons have become dependent on the finished structure of the program. Modularity: Free Software uses a modular approach to allow for unsurpassed development cycles - programs can be ported to other environments easily because the source is available for all components. This leads to popular programs being ported by developers who know the new environment intimately - instead of one team having to learn the new environment before porting.
It is not obvious to {me} that usability testing cannot be done with Open
Source software (initial inspection shows such a suggestion to be false),
with Free Software
nor that the reuslts of usability testing cannot be fed back into
nor that the results
improvements in the interface.
Open Source does not have to encourage contribution of new code back to the original interface - it's Free Software and the GPL that does that function.
It is also, if one follows the writings of Jakob Nielsen of www.alertbox.com fame, clear that closed source software still suffers from a dearth of usability testing and improvement. It is correct to identify this area as a challenge for the Open Source
for the Free Software community
community, and, given that all previous challenges have been met with serial success, it is reasonable to anticipate advances in it.
The more successful GNU/Linux becomes, the more developers become exposed to it, the more the ideas are discussed in high circulation media titles. The cycle brings fresh developers into the fold - people without the time to develop the entire project or without the desire to become shackled to a corporate software house with deadlines to meet. These people make the difference to Free Software by contributing in smaller ways that reflect an immensely more diverse test platform than any proprietary system can achieve. Bug reports, wishlists, mailing lists, documentation - vital components of any usable software - feed back into the project, inspiring further development for the benefit of the users. This contrasts strongly with proprietary houses where the only development incentive is to make it different enough to warrant a substantial investment in a new release. Releasing software without cost is a significant stimulus to Free Software development.
Innovation --------------- Most of programming and development effort, regardless of where it is done, falls into two large areas of work:- - Mundane: on the one hand there is the routine solving of well-understood problems in similar, identical, or slightly different circumstances to those in whcih they have been solved before. Closed source development
those in which Proprietary development has,
has, for instance, solved some of the the problems of local government IT support in such a way as to give "Approximately 16,000 justice and public safety-related data elements were collected from various local and state government sources. These were analyzed and reduced to around 2,000 unique data elements that were then incorporated into about 300 data objects or reusable components" (Adelman: http://consultingtimes.com/osgov.html ) - Particular: on the other there is the solving of a problem specific to a time and palce, a company or project, using what is to hand. THis may
time and place .... This may
involve inventing something new or achieving an insight, but this comes from the exact circumstances, and clearly is not made less likely by the source code's being visible elsewhere.
There are clear benefits when the source code is available and modifiable under Free Software that may not be present in Open Source. The sheer number and variety of developers willing to work under free licences like the GPL and MPL means that inventing something new because of new circumstances becomes more likely and happens in a quicker timeframe. Developers willing to take on such tasks are already familiar with the new environment and are keen to get the software to work for them. Innovation in software is clearly different from innovation in engineering or pharmaceuticals. Software evolves from prior software. The article itself explains how Richard Stallman had to achieve the creation of the compiler, the editor and the command shell before the licence and then the kernel could be created as Free Software. These were not clones of Unix versions, as the article asserts, because a clone insinuates an unbreakable inheritance link between the two - something that only SCO believe ever existed between GNU and Unix. The functionality of certain proprietary software has been reimplemented in Free Software but that is not the same as cloning proprietary software.
Of the remainder, some is speculative, some academic. Contrary to the assertion in the article, in the field of medical records software and medical ontology, the innovative projects are more likely to be Open Source (eg OpenGalen at Manchester University, Open-EHR in Australia). Each is very specifically engaged in developing "an independent software architecture" to support doing well things that are impossible or done badly at present.
Free Software is also well equipped to take on less popular architectures and allow the code to continue to develop, long after any commercial interest has died. This keeps legacy systems in operation, much to the annoyance of proprietary software houses with their intimate links to the hardware industry. Many proprietary hardware systems that have gone out of production are now available as open source or free software emulators. New software architectures are currently driven by the hardware manufacturers of peripheral items - PDA, camera, mobile, DVD etc. Very little is new within the architecture of the PC since the days of the 486. The challenge for Free Software and Open Source is to persuade peripheral hardware manufacturers to develop to the same mechanisms as system and network hardware manufacturers: open standards like ethernet, SCSI, ATAPI and others. With common, open, standards there is no reason why Free Software would not be the first to support new peripheral hardware architectures because development is usually faster and more efficient. Standards are the latest focus of tensions between proprietary corporates and Free Software. -- Neil Williams ============= http://www.codehelp.co.uk/ http://www.dclug.org.uk/ http://www.isbn.org.uk/ http://sourceforge.net/projects/isbnsearch/ http://www.biglumber.com/x/web?qs=0x8801094A28BCB3E3
Attachment:
pgp00067.pgp
Description: signature