Roy Schestowitz <newsgroups@xxxxxxxxxxxxxxx> writes:
> __/ [ Linonut ] on Friday 06 October 2006 01:01 \__
>
>> After takin' a swig o' grog, Ian belched out this bit o' wisdom:
>>
>>> .. that it nearly works.
>>>
>>> Unfortunately, nearly is not good enough.
>>>
>>> People spend mega man-years developing a project but in the end get
>>> tired or get a job and the project is never finished.
>>>
>>> Example. Look at Monodoc in Ubuntu. Expand the classes tree and what
>>> does the doco say ? 'the documentation for this section is yet to be
>>> entered'. That is, useless.
>>>
>>> Example Plone. Works with a version of Zope about 2 years old which has
>>> to be downloaded and compiled. Try the tutorial. Doesn't work until you
>>> hunt around old newsgroups and find that this has been a problem forever
>>> and new users are supposed to know that an id can't contain spaces.
>>>
>>> help very rarely is present or relevant.
>>>
>>> Someone has to take charge ! Redhat or someone has to lay down rules or
>>> guidelines that everyone adheres to,
>>
>> No way.
>>
>>> In my experience, computer sites of organisations using Linux/Unix are
>>> in a continual state of semi-chaos as they need systems programmers to
>>> be continually patching the site.
>>
>> You only offered two failures. There are far more than that.
>>
>> Yet, you are wrong. There are also a large number of thriving projects,
>> and they thrive because of lot of people love the software and use
>> it.
>
> Precisely. The worst model of software development is one which involved a
> preset scedule (deadlines), a set of requirements that cannot be bent
> and
More clues that Roy has never worked in the real world. Witness OSS
application SW : a lot of it abandoned. A lot of it half working but
feature rich. Some very good. Most of the really good stuff concentrates
on one or two things and does them well.
Deadlines are required in the real world so that parallel developments
can meet at MILESTONES. Not everything is just code and release Roy. See
your "free search engine" for a good example of how things can never get
done without suitable resource allocation and a decent project plan in place.
http://www.iuron.com
Imagine someone had put all their balls into one court and based their
future project on you delivering that? Whoops. Another SW development
team for the grim reaper.
> management layer (let the 'geeks' manage one another). Evidence on the
Geeks cant. It causes feature creep. And feature creep means nothing
ever gets finished. And most geeks are totally unable to appraise their
own SW and its UIs - as geeks they know best. Unfortunately, normally
they don't.
> surface appears to support this. Look at WordPress and Apache for example...
>
> Best wishes,
>
> Roy
>
> PSh - I love OpenOffice (version 2). It just works. I just don't use it very
> often.
Then how do you know it works?
You're a fraud.
--
Dijkstra probably hates me
(Linus Torvalds, in kernel/sched.c)
|
|
