In comp.os.linux.advocacy, Jimserac
<Jimserac@xxxxxxxxx>
wrote
on Mon, 08 Oct 2007 15:35:47 -0000
<1191857747.808608.208870@xxxxxxxxxxxxxxxxxxxxxxxxxxx>:
> On Oct 8, 6:33 am, Roy Schestowitz <newsgro...@xxxxxxxxxxxxxxx> wrote:
[snip of Roy's stuff for brevity]
> FUCK Microsoft Windows, I'm done with them quite some time ago.
I wish I could say the same for all of us.
1. A brain-dead application that assume certain
IE-specific cookie behavior is still used at $EMPLOYER;
it simply refuses to function with Mozilla. (Something
to do with cookies.)
2. The EU-required 'N' series is an interesting testimonial
to the fact that Microsoft is dominant in the media
player industry. Presumably, there are other solutions
such as RealAudio, but not that many.
3. Various artifacts in Gentoo point to Microsoft
dominance. The most important of these are certain
fonts and codecs. It is nice that Microsoft provided
them for free for limited personal use, but one
wonders why.
4. The WinE project is an interesting holdover, and
works very effectively. It is far from clear,
however, that in an ideal environment WinE would
even get started; the general idea instead would be
to define a common widget set that everyone can use.
(There are many issues with such a set, of course;
currently we have Motif, Qt, and Gtk. Of these I
prefer Gtk, mostly because it's a cleaner, newer design;
however, all three are still in use. The Windows one --
Win32 -- is a fourth candidate in this list, of course.)
5. Bodge-ups and workarounds in Websites are common
for that ugly monstrosity of a browser, IE 6
(IE7 might fix some of them, but botch up others).
The most notable problem is the differences in behavior
regarding AJAX; fortunately, these are confined to the
launch module. However, certain other Javascript/DOM
issues are not.
The most *visible* problem is arguably PNG transparency,
which is finally (apparently) fixed in IE7. Wonder
what took them so long? There are other issues in
CSS and XSL that surface from time to time, as well;
one ugly one is that, if Apache has .xml, IE tends to
interpret it as an XML file rather than an XHTML file.
Fortunately, IE works more or less correctly if one
specifies an XSL stylesheet within the XML file,
and one might argue that one should use document()
within the XSL anyway, and a stub XML referring to
the actual data using an attribute and xlink:href
(xmlns:xlink="http://www.w3.org/1999/xlink"), for a
more flexible design.
6. <APPLET> is officially dead; it has been replaced
by <OBJECT>, and Microsoft couldn't even implement
*that* correctly, forcing Sun to use rather incoherent
Microsoft-compliant class IDS. In a more ideal world,
something like
<OBJECT codebase="jars" archive="a.jar;b.jar"
classid="java:com.myCompany.anApplet.Main"
data="mydata.stuff"
standby="Loading applet..."
height="..." width="...">
</OBJECT>
would stand a chance of working, and
the applet could even pick up the @data
attribute. But no... classid apparently must be
"clsid:8AD9C840-044E-11D1-B3E9-00805F499D93", which
makes no sense to anybody. BOTCH!
7. It is not clear how standard
<html><body><svg ...> ... </svg> ... </body></html>
is, but the gyrations getting the Adobe plugin to
work properly within <OBJECT> aren't all that pretty.
Even Mozzie is less than happy therewith. Ideally,
<img src="#svg">
or
<object data="#svgimg" type="image/xml+svg">
...
</object>
...
<svg id="svgimg">
would just work fine; it would simplify things
a lot for certain types of websites. (Not sure
regarding embedded binaries, admittedly.)
8. Another monstrosity that surfaced during $EMPLOYER's
testing, caused by IE. Apparently IE does a socket
reset rather than waiting around for a socket close.
This is most clearly seen by looking through tcpdump or
other such output, if one has Linux configured to sniff
the network while routing a Windows machine through it.
One can also see such packets if one has installed IE
on one's Linux box using WinE and/or IES4Linux.
This is happening even now, on my one work box:
09:27:59.390804 IP h-x-x-x-x.desktops.x.com.37700 >
h-x-x-x-x.desktops.x.com.http: R 1068475766:1068475766(0)
win 0
^^^ (!! ideally should be F)
This is not quite a violation of the standard, and
their intentions were good (the idea is to reduce
shutdown time on their end of the socket connection),
but surprised the hell out of everybody here once
we discovered it, and one might consider it rather
impolite. Fortunately, Apache largely ignores such
transgressions, if it sees them at all.
9. Most OEMs simply preinstall Windows (once they've
debugged it sufficiently so that they can plaster
images on their drives during system build and burn in).
Until Linux can get penetration into the preinstallation
space (some vendors already offer preinstalled Linux,
others are dipping their toes in, and ASUS is taking
a very interesting leap which we'll see the details
October 10th but looks very exciting already), we're
stuck with Microsoft by default, and Microsoft gets paid
whether we really want Windows or not. Great racket
for them, terrible for the end consumer.
10. Application developers might simply assume Windows (or
develop first on Windows, giving it priority).
The results for Linux are less than ideal, though if
the app developers have done their homework (kudos
to those that developed Quake, Unreal Tournament and
other such, for example) it's a simple port to Linux.
11. Spamware, anybody? While the exact causes are unknown,
my email box gets a *lot* of crap (both home and work).
The general suggestion is that such is caused by
zombiebot networks, running infected versions of
Microsoft Windows. Of lesser concern (since NAT is
becoming commonplace for home networks, apparently)
is that probes still come in with malicious intent
on various well-known ports; these probes are, well,
probing for vulnerabilities they can exploit. Three
guesses where these probes are coming from.
12. OOXML is standard, in many senses of the word (but
not the proper one).
- It's widely prevalent, or will be once MS Office
2007 becomes installed on more business machines.
- It's the expected norm for XML documentation
produced by Microsoft Office 2007.
and of course
- It's a 6000+-page piece of utter crap, from all accounts.
We'll see what happens. Presumably Microsoft will
send them a revision or two that ISO will happily --
or at least grudgingly -- accept.
http://www.pcworld.com/article/id,136711-c,techindustrytrends/article.html
13. Sony's DRM rootkit looks set to do a bit of damage,
not because Sony released it (and botched it), but
because it's easily modifiable to generate certain
stealth viruses. Even those that favor DRM can't
like this result.
14. Apart from Exxon-Mobil and other such, Microsoft is one
of the biggest, richest companies in the world -- and
is arguably the biggest, richest software-based one.
Think on that for awhile (and then ponder what went
so wrong).
Granted, Forbes ranks them #55, as of early last year:
http://www.forbes.com/lists/2006/18/06f2000_The-Forbes-2000_Rank.html
It turns out Novell is #1800, and RedHat isn't even
on the list.
15. The DirectX versus OpenGL debate is an old one, and I'm
not sure which is better, except that OpenGL runs on
most platforms, whereas DirectX works well on exactly
two (Win2k and XP), might work on one more (Win9x),
and might function on one more (Vista), though WinE
might allow it to work on Linux/x86.
16. ActiveX. Blecch. 'Nuff said.
17. J++. Double blecch.
18. C#/.NET. So far, the adoption of this "standard"
(and at least part of it *is* a standard, routed
through ECMA -- but only part) has been characterized
as anything from slow to glacial to "molasses in
January left outside during a week-long blizzard".
Nevertheless, some websites are in fact using it.
Welcome to the New World Order. Where did you want to be led today?
>
> Citizen Jimserac
>
--
#191, ewill3@xxxxxxxxxxxxx
Linux. The choice of a GNU generation.
Windows. The choice of a bunch of people who like very weird behavior on
a regular basis, random crashes, and "extend, embrace, and extinguish".
--
Posted via a free Usenet account from http://www.teranews.com
|
|