chopping and changing ...

[Michael Meeks <michael ximian com>]

Hi Maciej,

On Mon, 2002-09-02 at 19:06, Maciej Stachowiak wrote:
> Apps can be, and indeed have been, obsoleted by a new better app in
> the same category, although there may be a period when both must be
> supported. Examples of past, current and future transitions of this
> sort include

	None of the examples you give happened in a corner; and I'm well aware
of them. And yes - traditionally Linux distributions have been able to
simply chop and change at a moment's notice.

	However - my thesis, which was clearly not stated explicitely enough is
that creating productized software is far more expensive than simply
changing the emphasis between several broken alternatives.

	To whit; the documentation alone is as expensive as the code - if
properly translated, similarly making things accessible can impose quite
a cost; then the better known costs of i18n, multi-head, and so on also
increase the cost.

	Thus 'just' switching to a different app needs to be based on an
extremely informed decision. Furthermore, if we have several million
deployed seats, there has to be a perfectly smooth upgrade path.

	All of this miltates against switching to a new application, and in
favour of maintaining the old. Thus it seems to me increasingly
important - as Gnome gains market share, and serious commercial support
- that we eschew the "hey we'll just switch some modules in core" in
favour of a more public, accountable, open, protracted process.
Furthermore, it seems to me that since we'll have to be maintaining
these things for the concievable future - we should pay careful
attention to what we're letting ourselves in for, especially if we are
blessing them as part of the Gnome core.

	I hope that clarifies my rational,

	Regards,

		Michael.

-- 
 mmeeks gnu org  <><, Pseudo Engineer, itinerant idiot