Mailinglist Archive: opensuse (4348 mails)
| < Previous | Next > |
Re: [SLE] An interesting review
- From: dep <dep@xxxxxxxxxxxxxxxx>
- Date: Tue, 8 Oct 2002 07:05:15 -0400
- Message-id: <200210080705.15480.dep@xxxxxxxxxxxxxxxx>
begin Ben Rosenberg's quote:
| This woman reviews based on how things look based esthetics and as
| a slightly smarter then average dumb end user. And if you dispute
| anything she says via talk backs..she loses it. She told a user the
| other day to f**k himself and was quite mean to him. Just because
| he begged to differ on her opinion. I wouldn't give her the click
| though.
forgive my saying so, but without looking directly at the accuracy of
your appraisal, which might be spot on, reviews of distributions
oughtn't be looked at individually, but instead multiple reviews
taken together can give a pretty good sense of a product. there are a
couple reasons for this, but the most important is that a modern
distribution is a huge thing that cannot be comprehensively explored
under deadline pressure, and anyway what's important to some is
insignificant to others. for instance, a lot of people make a big
deal out of the installer, while others are willing to endure hell's
own installation if the result is a rock-solid system. some
concentrate on robust package management, while others hold the view
that the thing ain't installed until the rpm database is broken. some
concentrate on aesthetics (and not just slightly smarter than average
end users -- look at the dispute between kde and red hat), and there
is argument in support of this, because after all that's what the
user deals with. otoh, there are those who couldn't care less,
because they blow away whatever's shipped in favor of their own
customizations anyway, or else they eschew X and desktops entirely.
there's just no way that a single review can take all of this into
account. the attempts to quantify reviews through performance tests
have invariably fallen short because all that does is move the
subjective nature of reviews back one notch, to decide which
performance features are important and therefore worth testing. even
hardware tests suffer from this: ideally, they would in aggregate
tell how quickly one can do stuff, but again the problem -- what
stuff?
so the answer is to read as many reviews as possible, if one is
interested in that sort of advice, picking and choosing among the
things that are important to the reader.
--
dep
http://www.linuxandmain.com -- outside the box, barely within the
envelope, and no animated paperclip anywhere.
| This woman reviews based on how things look based esthetics and as
| a slightly smarter then average dumb end user. And if you dispute
| anything she says via talk backs..she loses it. She told a user the
| other day to f**k himself and was quite mean to him. Just because
| he begged to differ on her opinion. I wouldn't give her the click
| though.
forgive my saying so, but without looking directly at the accuracy of
your appraisal, which might be spot on, reviews of distributions
oughtn't be looked at individually, but instead multiple reviews
taken together can give a pretty good sense of a product. there are a
couple reasons for this, but the most important is that a modern
distribution is a huge thing that cannot be comprehensively explored
under deadline pressure, and anyway what's important to some is
insignificant to others. for instance, a lot of people make a big
deal out of the installer, while others are willing to endure hell's
own installation if the result is a rock-solid system. some
concentrate on robust package management, while others hold the view
that the thing ain't installed until the rpm database is broken. some
concentrate on aesthetics (and not just slightly smarter than average
end users -- look at the dispute between kde and red hat), and there
is argument in support of this, because after all that's what the
user deals with. otoh, there are those who couldn't care less,
because they blow away whatever's shipped in favor of their own
customizations anyway, or else they eschew X and desktops entirely.
there's just no way that a single review can take all of this into
account. the attempts to quantify reviews through performance tests
have invariably fallen short because all that does is move the
subjective nature of reviews back one notch, to decide which
performance features are important and therefore worth testing. even
hardware tests suffer from this: ideally, they would in aggregate
tell how quickly one can do stuff, but again the problem -- what
stuff?
so the answer is to read as many reviews as possible, if one is
interested in that sort of advice, picking and choosing among the
things that are important to the reader.
--
dep
http://www.linuxandmain.com -- outside the box, barely within the
envelope, and no animated paperclip anywhere.
| < Previous | Next > |