Fragmentation irritates me.

13 Jun 2004

But what the hell, I’ll respond here, anyway.  Dale, hurry up and get a commenting system!

In response to the entry on his front page (as of 13/06/04 – no permalink available), entitled “The truth behind open source, closed source and anything in between.”

Specifically, in response to comments made with regard to the Fedora distribution of the Linux OS; Fedora, whilst officially a technology release, holds technological advancement as a priority (objective #5) – this, if not adequately tested, may result in instability and a product which is NOT usable by the wider community (hence in conflict with objective #1).

This is something of an occasional neccessity – they are (so it would seem) locked down by their incredibly tight development schedule, as a result of this being an established requirement of the Fedora team (presumably in order to permit Red Hat to fully evaluate technologies in advance, with a wide-testing audience, in order to ensure the continued quality of its’ commercial offerings).

That said, it is perfectly reasonable for Red Hat to NOT offer software support for this distribution – such is the nature of Open Source.  It is not financially viable for Red Hat to offer [any formal level of] support to users of Fedora, as Fedora is a technology release, released free of charge to the community.  This IS very much beta-quality software.  I am sitting here on a FC1 desktop – I have been using this as my primary operating environment since the beginning of the year, and Red Hat 9 before then.  It is usable, in the same way any Beta-quality software is usable.  Rough edges, occasionally, but not typical of the finished product (or the development model!).

In the same way the Apache organisation co-ordinates the development of its various product offerings (we’ll stick to HTTPD for the sake of argument), Red Hat and other distribution development organisations are similarly responsible for the co-ordination of their respective developments.

The fundamental difference between Apache and Red Hat is the nature of their “flagship” product.  Apache doesn’t have a flagship product (they are best known for their  HTTPD engine, but if you visit http://www.apache.org it is evident that there is a wider range of software available and being developed under the Apache product, all of which is equally promoted, regardless as to market share/penetration or apparent application), whilst Red Hat do.  RHEL carries SLA’s, stability, and a price-tag.  Apache HTTPD carries none of the above, and functions perfectly well (on something like 60% of the worlds’ publicly-accessible web servers).

So what is “different” (and therefore, presumably, wrong) about the development model employed by Red Hat?  To put it bluntly, they are attempting to balance priorities of bringing a commercial product to market, whilst simultaneously attempting to provide a completely open derivative within established time constaints.  Conceptually, Fedora is supposed to be a software breeding ground, for integration into RHEL when it reaches maturity.  Realistically?  Fedora is a testing ground.  They do release products to schedule (I would argue that this is their failing point).  And components do eventually become integrated into RHEL products as they reach maturity.

What was wrong with FC2?  I don’t know – I haven’t used it (haha, I have a GB or two of ISO’s sitting there still un-burnt).  In a comparison such as the one Dale claims to be making, I strongly object to any complaints made about the boot-loader issue.  When you show me ANY Windows OS boot-loaders that will support a non-MS platform (heck, one that didn’t overwrite the MBR by default would do me fine!), I’ll heed your compliants.  Until then, write to Microsoft about it asking them to resolve the problem.

Because you can’t modify that part of the Windows installer yourself.  Something to do with closed source concepts, perhaps?