backtop

Print E-mail del.icio.us 134 comment(s) - last by kayronjm.. on Apr 17 at 5:26 PM
The truth comes out about User Account Control

Microsoft's Windows Vista operating system has been lambasted ever since it was launched for consumers in January 2007. Diehard Windows users balked at the steep system requirements, sometimes sluggish performance, inadequate driver support, and varying products SKUs at multiple price points.

One feature that has caused quite a bit of controversy with consumers has been the User Account Control (UAC) that is included in Windows Vista. UAC prompts nag users for simple operations such as going to device manager, emptying the recycle bin, or installing/uninstalling an application.

David Cross, a product manager responsible for designing UAC, gave the real reason for UAC at the RSA 2008 conference in San Francisco yesterday. "The reason we put UAC into the platform was to annoy users. I'm serious," remarked Cross.

Cross added that Microsoft's unorthodox method to stop users from wreaking havoc with their systems and to stop software makers from making applications that delved too far into the Windows subsystem was a necessary move.

"We needed to change the ecosystem, and we needed a heavy hammer to do it," Cross added. Cross went on to say that although UAC may be seen as an annoyance to some, but its lasting implications are far more beneficial to Vista users. "Most users, on a daily basis, actually have zero UAC prompts."

Many would say that many users have zero UAC prompts on a daily basis because they have already disabled UAC -- not so says Microsoft. According to Cross, 88% of Vista users have UAC enabled and 66% of Windows sessions do not encounter UAC prompts.

Comments     Threshold


This article is over a month old, voting and posting comments is disabled
Sluggish Performance
By UppityMatt on 4/11/2008 11:10:25 AM , Rating: 4
This is just my thoughts here so don't go crazy down rating me or anything. Does anyone else get the impression that the more our hardware advances the less programmers worry about actually shrinking and making code more effective. I have taken several programming classes in college and the professors would always mark us down if we used a double instead of an INT when we were suppose to. It almost seems like alot of programmers that actually do it for a living just get the job done and don't worry about optimizing. Vista to me seems to fit this bill, i know they crammed alot into it but i think they could have easily worked more on optimizing.




RE: Sluggish Performance
By Chris Peredun on 4/11/2008 11:25:25 AM , Rating: 5
quote:
Does anyone else get the impression that the more our hardware advances the less programmers worry about actually shrinking and making code more effective?


Yes, especially in the learning environment. XNA isn't helping that matter much - from my strolling around the XNA Developers forum, far too many people seem content to use horrifically inefficient methods and letting the PC/360 grind its way through them.

Though I would love to see their heads explode if you asked them to code something for an ARM platform without floating-point; say, the Nintendo DS.


RE: Sluggish Performance
By Mitch101 on 4/11/2008 3:32:53 PM , Rating: 2
Any of you played the game Transportation Tycoon when it was a DOS based game? In its game format back in the dos days? Chris Sawyer should have taken a shot at making the game into an OS. Multitasking in that game was excellent and having the many windows open tracking various items. It took a lot to slow that game down and that was in the 66Mhz cpu days. Plus the game never crashed. I think the original game fit on a floppy too.

That game had me amazed at how well he had programmed that game.

The game did in the end have a bug that he must have not used double for the cash amount because you could exceed the cash limit late in the game. It didn't crash but your money would turn from positive to negative billions. Just one bit!. Still awesome programming at its best.

Most code is a mess today. Calling the same objects multiple times instead of once just kills me. Why didn't he get all the elements the first time he connected to the database? Why so many calls? Grrr!


RE: Sluggish Performance
By smitty3268 on 4/11/2008 10:30:40 PM , Rating: 2
No need to use a double, just a long. Or at least an unsigned int which would have doubled the limit and caused the overflow to go back to a more reasonable 0 rather than -2 billion.

Was just reading this blog which was also talking about repeating the same code over and over again.

http://www.virtualdub.org/blog/pivot/entry.php?id=...


RE: Sluggish Performance
By BladeVenom on 4/11/2008 7:29:14 PM , Rating: 5
What Intel giveth, Microsoft taketh away.


RE: Sluggish Performance
By leidegre on 4/12/2008 5:55:23 AM , Rating: 4
I'll have to disagree, there are great talented people out there, but the bar is not as high as it used to be.

Yet, I can not just ignore the fact that people seem to waste computational power, a lot. As a developer myself I'm very picky about a particular software. I expect optimal performance and if that is not the case, I go otherwise.

The worst lie of them all is that we seem to have so much computational power that it doesn't matter any more. That's just out right stupid.


RE: Sluggish Performance
By Funksultan on 4/14/2008 7:45:13 AM , Rating: 4
Well, there is a ripple effect that becomes important here...

The purpose of marking you down is so that you truly understand the difference between int and long, not because it is going to effect the performance of your immediate program, but because it's important for you to know. An easy analogy is ebonics. Yes, we all can understand slang, but if we use ti to the point where we forget the use/spelling of the proper words, we lose something.

In the same way... yes, my quad-core is gonna chew up any code I write for it, pretty much no matter how sloppy I am, so I suppose I don't have to worry about long vs int. I might go years not caring, and eventually forget about ints all together.

What happens when I'm an OS programmer? Or what happens when I've trained 2 generations of junior programmers, and nobody uses ints anymore? Now, instead of a few occurances, the problem starts to magnify.

A coder who knows how things work vs. one who doesn't is like the difference between an architect, and someone nailing boards together. (yes, a Costanza moment)


RE: Sluggish Performance
By darkpaw on 4/11/2008 11:31:57 AM , Rating: 5
I would have to agree with you there. People that learned in a lower level language where they had to manually manage all of their memory allocations (C, Assembly) tend to use the smallest they need. People that learned something like Visual Basic in school tend to just assign the biggest they can so they don't have to worry about size.

I did a lot of embedded/C type stuff in my undergrad and learned good memory management practices then. It really seems the focus on current development for general purpose systems is just get it coded as fast as possible, memory footprint doesn't matter.

I like Vista and don't mind the memory requirements at all, it's not like 2gb is expensive. I do worry about where this trend is going though.


RE: Sluggish Performance
By murphyslabrat on 4/11/2008 12:31:55 PM , Rating: 2
quote:
I like Vista and don't mind the memory requirements at all, it's not like 2gb is expensive. I do worry about where this trend is going though.

You have Vista to thank for that, you know. ^_^

You can get a 2GB S0-DIMM now for $35 at the egg.


RE: Sluggish Performance
By ImSpartacus on 4/11/2008 3:41:52 PM , Rating: 3
2x2Gb is actually cheaper per GB than 2x1GB (it wasn't always that way). I say bargain bin A-Data 2x2GB for $50 after a $10ish rebate.

Who needs to pay 200$ for 2GB of DDR3 when $50 can get you similar performance and double the capacity?

It wasn't always this way, but Vista is certainly becoming manageable with current prices.


RE: Sluggish Performance
By mindless1 on 4/11/2008 9:32:40 PM , Rating: 2
I think you are missing the point, which is not memory cost but that at any given memory bandwidth it still takes longer to shuffle that code around. We're effectively cancelling performance gains and user productivity, the truth is people can't do the most common things on a PC any faster than they could 10 years ago (assuming certain things abandoned like floppy discs, dialup internet access) while in any other activity the user would certianly be faster after years of practice and using equipment potentially > 10X as fast.


RE: Sluggish Performance
By goku on 4/12/2008 5:16:23 AM , Rating: 2
I couldn't agree more.


RE: Sluggish Performance
By boogle on 4/11/2008 11:46:59 AM , Rating: 5
Using physically smaller datatypes is rarely a performance optimisation, its a memory optimisation. For example if the CPU reads in 32bit chunks (like all 32bit Intel CPUs), then using a 16bit short will not improve performance whatsoever. Storing 2 shorts in a single int would give a nice boost though (reading in two variables at once). Also floats (I know, I know, not a double) are just as fast as ints - it's the conversion process to/from int that is slow.

The performance issues aren't trivial like this, they arrise from the various utilities and libraries in use. For example rather than storing the result of a library call, it is called multiple times from within the very same method. Or using a generic object instead of a specific datatype (which in .NET/Java results in LOADS of boxing/unboxing), and so on.

The answer is to realise the cost of every method call, and program accordingly. .NET especially allows you to either write code almost as fast as C++, or write code that takes an eternity. This is what seperates a good programmer from a bad one - but given the current shortage of developers, bad programmers are all over the place. Want easy money? Graduate and program .NET.

But either way, genuine optimisation is an extremely complex subject. Deadlines are short. To say 'optimise' is significantly easier said than done. If we went back to 'real' programming, Windows NT would be a pipe dream and Windows 3 would only just be finishing development.


RE: Sluggish Performance
By Locutus465 on 4/11/2008 11:56:42 AM , Rating: 3
Actually quite the oppisit... Useing data types smaller (or wider) than the bit width of the processor will in many cases result in a preformance hit beleive it or not.


RE: Sluggish Performance
By boogle on 4/11/2008 12:00:04 PM , Rating: 2
quote:
Actually quite the oppisit... Useing data types smaller (or wider) than the bit width of the processor will in many cases result in a preformance hit beleive it or not.


No no, I completely agree. However, you can get great speed improvements by putting lots of values into a single large stream (aka. using MMX/SSE) and working on them all at once.


RE: Sluggish Performance
By Locutus465 on 4/11/2008 12:35:50 PM , Rating: 4
that takes a lot of optimisation (fortunetly compilers do a lot of the heavy lifting these days for normal folk), and honestly for an OS I'm not so sure you would even see a difference. Addtionally you only get a gain if you're performing the same operation on all data values. It terms of Window's "bloated" state, I wouldn't look to carelessness being the reason why their OSs are getting heavier. Probably more a desier to increase the number of built in features (and running a good portion of them by default) as to being the reason why Windows OS code is getting heavier.

Just think, vista has:

Antispyware
Automated defragmetation
Pretty 3d ui
gadgets
Media center (home prem+)
Tablet PC functions
Extended networking featuers
and a host of other features you either had to get third party or buy a specialized version of windows to get.


By KristopherKubicki (blog) on 4/11/2008 12:00:09 PM , Rating: 2
He's right.


RE: Sluggish Performance
By Locutus465 on 4/11/2008 11:53:24 AM , Rating: 3
The answer in general commercial software, particularly with web based systems is yes... One would hope that Microsoft is designing their systems more efficiently though (if not packing with thousands of new features)... I'm guessing the quality of Microsoft's code is good since the last time Windows OS code was leaked into the open it was met with very good reviews (Windows 2000).


RE: Sluggish Performance
By PAPutzback on 4/11/2008 11:59:02 AM , Rating: 2
I think you are comparing different types of coders. I imagine on a new release there is quite a bit of code that wasn't optimized in order to get the product out the door with the plan for the service packs to supply the optimizations.

And a lot of the programmers who do it for a living like myself have so many pending requests you have to spit out code so fast to keep up or find a different living.

A lot of the time your managers might not have coded a day in their life so they have no clue what goes into program. I have had reports take longer to program to get what the user is expecting than a large SSIS package.

And then you have MS spitting out a new platform every six months implemented only half way. They gave us VS 2008 and it doesn't work with reporting services so now you have to maintain two VS installs or run a VMC.


RE: Sluggish Performance
By Locutus465 on 4/11/2008 1:53:22 PM , Rating: 2
Hows the stability of 2008? I've been hesitent to even consider giving it a try considering how flaky 2005 is. VS.Net was my all time fav environment in the 2002/2003 days, starting with 2005 I'd have to say it still integrates all the tools better than anything else I've used, but daaaaaaaaaamn does it sacrafice reliability.


RE: Sluggish Performance