__/ [ [H]omer ] on Monday 25 September 2006 04:35 \__
> .----
> | My customer are getting slowdown problems after installing WGA
> | Notifications - I am a self-employed technician who works on about
> | 1000 different computers a year. In the last month, I've ran into a
> | handful of customer who had a mysterious problem immediately after
> | installing the WGA notifications update.
> |
> | In each case, the customer would report to me that they computer was
> | dreadfully slow. Upon arriving on-site, I would find that a
> | particular process in task manager was maxing out the cpu. For each
> | customer this process was different. For one, it was svchost.exe, for
> | another it was qbw.exe (Quickbooks), and various other processes for
> | different customer. In each case, restarting the computer had no
> | effect. As soon as the offending process would start (in case of
> | Quickbooks, when they launched the program. In the case of
> | svchost.exe, immediately when the desktop would appear.), it would
> | take up 99% of CPU and stay that way forever. Reinstalling the
> | offending application wouldn't fix it.
> |
> | In each of these cases, the solution was to use system restore to go
> | back to before they install WGA Notifcations, and then reinstall all
> | updates EXCEPT wga notifications.
> |
> | In every case, my customer had legitimate OEM copy of Windows. Not
> | that it matters, but each was a Dell.
> |
> | In a way, I'm not upset that Microsoft rolled out this as an update.
> | My main competitor for years has been installing counterfeit copies
> | of Windows and Office, and now he must face the rath of unsuspecting
> | customers who get nagged every morning about their illegal software.
> |
> | However, I am looking for an admission from Microsoft that this
> | update does indeed cause this problem, and a downloaded fix for the
> | issue.
> `----
>
> - http://slated.org/wga_slows_down_xp
Having just gone to your Web site, I realise that the story is brought to us
from Microsoft's own turf. The guy/gal does not sound too pleased, at the
same time failing to comprehend that each patch/addon is a merely patch
rather than a reconstruction. It is bound to cause confusion. Each patch and
Service Pack is not built from the ground up. There are too many
dependencies. Simplified example:
,----[ Pseudo-ish code ]
| Server.checkgenuine()
| Server.connect()
`----
Now Microsoft realises that some site can actually pretend to be a Microsoft
server, which makes the validity test invalid (a bit of a pun). No
worries...
,----[ Pseudo-ish code ]
| Server.checkgenuine()
| if Server.isthisafake()
| break;
| end
| Server.connect()
`----
So rather than fixing checkgenuine(), you just patch using a condition. Maybe
there is too much risk involved with changing checkgenuine() on millions of
live installations (history suggests that patching and SP's breaks some
modules at times). All in all, such conditional statements and
non-principles approaches may be the reason why 60% of the Windows code
/still/ needs to be rewritten**.
Best wishes,
Roy
**Wrong model of abstraction aside, e.g. single-user with pseudo-multi-user
as an extension, which leads to flakiness in permissions and privileges
control, or escalation, arbitrary code execution with Admin permissions, for
after all, Windows still struggles to provide a fully-functional desktop to
non-Admins
--
Roy S. Schestowitz | #FFFFFFF4 ADD &R1, "9999999", &BankAccount
http://Schestowitz.com | Open Prospects ¦ PGP-Key: 0x74572E8E
Tasks: 148 total, 1 running, 144 sleeping, 0 stopped, 3 zombie
http://iuron.com - knowledge engine, not a search engine
|
|
