a b wrote:
1. Documentation is a major pain in the ass to find. Outside of man
pages and the occasional Sun engineer blog entry, there seems to be
no decent documentation. In fact, most people admit that the "Solaris
10" books that are currently out, are simply Solaris 9 books with a
new cover. How is a user who asks "is there a equivalent" supposed to
actually find out if it exists?
Personally, I find docs.sun.com to be an outstanding source of
documentation, especially on some very obscure technologies in Solaris.
Yes, there have been times when I wished that there were more
examples, and when I wished that the search engine was better, but
overall, docs.sun.com has top-notch content against which many other
computer companies can measure themselves. The documentation is
extremely well organized and structured. (docs.hp.com and
techpubs.sgi.com are comparable in that regard.)
You admit that docs.sun.com leaves something to desire. I'm not
advocating throwing everything out .... simply making it easier to use,
find and getting howtos and the like. I also understand that a huge part
of that is the community helping out ... and it'll get there, it's just
my opinion that the importance of documentation and the effort of
improving it needs to be at the forefront of everyone's mind.
2. Ease of use .... Now, I know that most of you old school UNIX guys
laugh at this, but usability is important. You've tuned me into a
cool way to do something along the lines of USE flags in Solaris, but
it sure sounds like it's not gonna be easy.
I guess the hardest part is getting the software to compile on
Solaris. After that it's just a matter of providing the correct
compile switches to Sun Studio compilers to generate corresponding
Instruction Set Architecture optimized binaries.
But other than that, how hard can it be to do a `cp` of
/usr/lib/isaexec to the name of the binary?
Simply doing a 'cp' is easy, unfortunately that's not the whole thing.
Actually getting the software to compile, getting your paths setup
correctly and getting it to use the right damn compiler takes a lot of
time and frustration that could simply just be done better.
Using Ubuntu for an example, the reason it's become so popular so
fast is not at all because it's superior. In fact, I can't stand
quite a few things about it. It's become so popular so fast, because
it's *EASY*. Without ease of use, no matter how awsome the feature
set is, you're just gonna end up being the thing people use only when
they absolutely have to.
Yes, but ease of use only works for desktop systems. Internet however
isn't powered by desktops, it's powered by a massive, decentralized
infrastructure, and is a very large number of businesses.
In practice, the convenience approach can only work so much, until the
whole model collapses. It might work very well for a single desktop
system, but after a certain people to systems ratio, the whole model
can't support the needs of the user or users any more.
Ultimately, neither the "convenience administration" nor a single
desktop system paradigm will be able to survive what will eventually
come our way. The writing's on the wall. Desktop's days are numbered;
it might take years, but it will come.
I completely disagree. There is a reason that Windows as a server is
where it is today. Ignoring other reasons, it's pretty easy to get setup
and running. In fact, in my opinion, that's one of the reasons that RH
is as popular as it is ..... it was easy to get up and running (of
course the support makes a big difference). This is especially important
when moving someone from system X to system Y. If I have to hack a
system to get it running the way my old architecture was, I generally
will just forget about it and go back to where I was.
_______________________________________________
opensolaris-discuss mailing list
[email protected]
