[Linux-aus] AT&T Tests Linux - boosting AT&T's leverage

Wed Oct 6 14:44:02 UTC 2004

Have to rant on this one...

The broken core of windows could be exactly the reason that LongHorn 
isn't being shoved out the doing post-haste as many previous Windows 
OS's have been. However, the big problem with this is still that if they 
want to support new-and-fancy-technology-X, some bugs are likely to 
creep in. From a standpoint of network security though, revamping the 
network model and removing the welded-tight integration between IE and 
the OS might help with a lot of problems related to worms, etc.

For those that read slashdot, there was an article about how programmers 
quite often get blamed for software malfunctions, when said malfuntion 
is directly related to lack of proper testing/debugging etc. And it's 
not that the programmers are totally at fault - in fact MS employees 
same damn good coders from what I hear - it's that when given the choice 
between glitter and security/testing, the glitter tends to creep into 
the project and push aside the testing. Again, linux suffers from this 
as well, but as many of the best linux projects are a collaboration of 
coders working toward a dream rather than marketers/PHBs working towards 
a budget (or making things fancy for a client), the end recent is often 
better.

And many linux projects do suck to begin with too, but they're also 
constantly in a state of update. Tools such as "apt-get" for debian 
ensure that coders can update a component or package and get the update 
out to the world. If a library gets updated and breaks something 
not-quite-right in the application, a new version of the application can 
be made to work with the updated lib.

Try doing that in a closed OS. While it seems to work fine for programs 
that need regular updates anyways (Norton, etc), could you imagine the 
average user running an update for PhotoShop, DreamWeaver, and 
RealPlayer? It won't happen, because it can't. While there is no 
relation between the applications, they can still impact upon each 
other. The cohesion between the applications and the OS isn't there 
until it hits your system. That's in addition to all the security 
issues, of course.

Right now my latest apt-get on debian unstable has broken one of my 
customized packages (IceWMCP which isn't available as a deb). I'd expect 
this to be fixed soon as new versions of the app/libs are regular. Won't 
see this in closed source.

LongHorn may solve some of these problems, but in some cases I don't see 
how. Can't re-write the OS from scratch and fix everything because might 
very well break legacy apps. If it won't run all your existing software, 
it's less useful than the alternatives (linux). Even if they manage it, 
the deadline for it to hit shelves is quite a ways away, and as we can 
see Linux is RAPIDLY gaining mindshare and even marketshare. MS is stuck 
right now. They've got an operating system that's broken at the core. 
They've got thousands of apps that scream for compatability. Continue 
with either broken, and they're screwed. Even it AT&T is flirting Linux 
to get a discount from MS, others will notice, and it will eat away at MS.

Still, there's nothing more dangerous than a trapped animal, and MS is a 
1000 pound gorilla trapped in a corner. If things don't look like 
they're fixable on the platform battleground, we'll either see two outcomes:
a) They switch to new, different types of innovation (and yes they've 
got lots of branches, X-box, cellular, etc) - but anti-trust might 
stymmy some of this
b) They focus more on litigation. Specifically patents, which can have 
the effects of damaging compatability for Linux, sullying our reputation 
(those damn OpenSource developers stole our ideas and infringed on our 
patents), and generally blocking/slowing development. We've already seen 
it starting, and - just like the windows viruses are multiplying faster 
than ever - so will idiotic patent issues.

Perhaps we need to start GNU-Patents or something... collect as many 
patents as we can based on OpenSource capabilities. Remember, while 
parents can be inherently evil - much of that evil is in their use.

I think that about ends my rant...

Leon Brooks wrote:

>Quoting:
>
>http://quote.bloomberg.com/apps/news?pid=10000103&sid=aZ2JnBlm5tOs
>
>  
>
>>"I still have concerns about security" in Windows, Eslambolchi said.
>>"We have had more viruses attacking PCs in the last six months than
>>in the previous 10 years." 
>>    
>>
>
>That's going to get worse, not better.
>
>  
>
>>"If Microsoft solves the security problem, and I think they will, I may
>>not have to switch," he said. 
>>    
>>
>
>The big insoluble problem at the base of all of this is that MS-Windows was 
>not design secure from Day One, whereas Linux (through a Unix-like conceptual 
>heritage) has been.
>
>Microsoft had a chance to fix that when they adopted VMS (as MICA) wholesale 
>for their NT stream of products (2000, XP, 2003, ShortHorn), but botched it 
>by working too hard for compatibility with their MSDOS-based horribly 
>insecure barely-timesharing 9X product stream. One single configuration 
>options gives VMS high-level military security, which you can only do for 
>MS-Windows by unplugging everything - no network, no I/O - and how useful is 
>that?
>
>In their current situation, they (and their users) are faced with an unending 
>game of whack-a-shark. It's like trying to pave over kikuyu grass - the 
>problem won't go away until you rip up the paving and get every shred of 
>kikuyu out first. The bullet that they failed to bite up front, they're 
>having to bite now, and they're foolishly trying to do it by paving over the 
>problems rather than solving them at their roots.
>
>The consequences for the user include more difficult-to-use versions of 
>Internet Explorer (which will also only get worse), Outlook and so on.
>
>The other problem Microsoft faces is monoculture. They only have one of 
>everything, not even a dichotomy, and for all practical purposes they're down 
>to one hardware platform too.
>
>Linux, by contrast, is a unity-in-diversity situation, with many varied 
>products able to freely interoperate because they all stick to the same 
>published standards rather than because they've all been regimented. This 
>means that users get more choices and still get to integrate stuff well 
>(Microsoft's current favourite feature) but they do it without laying out a 
>welcome mat for virus writers.
>
>  
>
>>Some companies may express interest in Linux to gain leverage in
>>contract negotiations with Microsoft, said Charles Di Bona, an analyst
>>at Sanford C. Bernstein & Co. in New York.
>>    
>>
>
>He's absolutely right - Newham in England being a classic example, and Telstra 
>being a local (for me) example - and from the outside looking in through this 
>one news report, that looks exactly like what AT&T is striving for.
>
>If so, there are some simple things AT&T can do to improve their bargaining 
>position, and they can be started right now.
>
> * Test all of AT&T's websites for compatibility with the Mozilla suite of
>   products, badge them as such, add the FireFox web browser to the software
>   AT&T rolls out to its PCs and internally promote its use; and
>
> * Add the ThunderBird email client as well, promoting it as a more secure
>   alternative to Outlook; and
>
> * Add the OpenOffice suite as well, setting the default save format to be the
>   Microsoft document types for now in order to minimise disruption of
>   existing workflow, and promoting that for its ability to recover broken
>   documents, its autosave which works reliably, and its ability to spit out
>   PDFs without additional software; and
>
> * Spam all of your employees with a free copy of TheOpenCD so they can
>   use these things at home as well, making it clear that this is free and
>   completely legal to copy and share.
>
>The nett effect of this will be to show Microsoft that AT&T are serious about 
>alternatives, to gain experience in using potential alternates, and to begin 
>preparing AT&T's people for a switch if they do elect to go ahead (all of the 
>above products run identically on MS-Windows, Mac OS X and Linux - what 
>training costs?). It can be done today and without ploughing up any existing 
>systems.
>
>The cost will be minimal (some extra bytes in the rollouts, pressing of 
>employee CDs at maybe 20c a pop and distribution of same) and even if AT&T 
>are really only fishing for discounts, it ups their leverage considerably.
>
>For example, OpenOffice does all of the day-to-day tasks that MS-Office does, 
>plus writes PDFs and uses a published XML-based standard (OASIS) for its 
>native documents - Microsoft are going to have to show significant positive 
>value in their product just to bring it up to $0 worth!
>
>Cheers; Leon
>
>
>References:
>FireFox - http://www.mozilla.org/products/firefox/
>ThunderBird - http://www.mozilla.org/products/thunderbird/
>OpenOffice - http://www.openoffice.org/
>TheOpenCD - http://theopencd.sunsite.dk/
>The OASIS file format standard - http://www.oasis-open.org/
>
>_______________________________________________
>linux-aus mailing list
>linux-aus at lists.linux.org.au
>http://lists.linux.org.au/listinfo/linux-aus
>
>
>  
>