INQ hopelessly biased against Microsoft

Hi

I thought I would use the wonder of free email to complain about theinquirer.net's continuous and obviously biased coverage of all things Microsoft.

Every opportunity seems to be taken to run-down Microsoft and extol the virtues of Linux. In fact I am at a loss to think of any article that I have ever read on your Website which is anything but derogatory to Bill Gates and his marvelous invention.

I appreciate that Microsoft products can cause problems and that there are numerous security issues and patches that need applying, but considering the sheer volume of computers running the software across the world and the number of hackers who dedicate their lives to finding vulnerabilities, I don't really find it surprising - even with this in mind, patches are automatically installed at the click of a mouse.

On the subject of your much loved and praised Linux, I have tried on several occasions to run a Linux server or workstation and to no avail (I am an IT manager). Drivers are impossible to find and instructions take no account of what happens when the install deviates from the expected path. It may be stable, but that's sod all use if I can't get it onto a machine. On top of this, with Windows, information is laughingly easy to find when you are having problems, and software of all shapes and sizes is easily, downloaded or bought and installed. Linux on the other hand is hard to find information on because there are highly differing opinions from a multitude of Linux wannabe pro's and a ridiculous number of people ignore or make fun of "N00b" questions.

I think that you should give Microsoft at least some respect for providing a product that a plethora of people are happy to use, and that give non-technical persons an easy, out of the box route to the Internet. Linux - sure, it has its place, but so does Windows.

Anon Fankoo

Disregarding the fact that buffer overruns are a fact of life in the Microsoft OSes to complain about bias on the part of Inq is ludicrous, ignorant and without any merit whatsoever. Those making the argument that if other OSes were to receive the same scrutiny they would also fail the same way simply have no concept of what's involved.

For example: It is not possible to perform a buffer overrun in the Pick environment. The OS doesn't allow it. This has been true since 1977. Preventing buffer overruns in AS400, Pick and VMS is part of the basic OS architecture. The science behind such architectures is well understood by all with the obvious exception of Microsoft. Even *nix people understand it. Witness the invention of and implementation of stack protection in various *nix kernels. Stack protection isn't common yet. But it's available now and shows great promise. It's another example showing where Microsoft has yet to catch up in order to equal the technology of true innovators.

To all the Microsofties out there looking to defend the Redmond Behemouth. I have no problem with where your loyalties lie or your choice to defend Microsoft. But try to understand what you're talking about first. Hiding your head in the sand by ignoring Microsoft's serious shortcomings or claiming that other OSes share the same problems when they obviously don't does yourself, your emplorers and your customer or clients no good.

People who think M$ software is mediocre don't know the half of it.

Name supplied

Hi...

Of course, even if Windows doesn't use the gimmicks which are offered by the memory management hardware, it seems Microsoft still hasn't managed to hammer good programming practices into the brains of their programmers. Coping with buffer lengths is, as a friend of mine would call it, a no-brainer. Usually all APIs where buffers are needed, also allow specification of the length of that buffer. All communications protocols usually use length fields, so it is not really advanced magic to write software that can handle insufficient buffer space.

Maybe there is some money to be made from installing programmer boot camps and making them as hard as US Marine Corps training, just hammer good programming practices into the brains of the recruits. Some bald-headed 250 pound Sergeant blasting buffer handling into their brains at 98dBs, while they're doing sit-ups could have the desired effect sooner or later.

Best regards,

Michael

Mike,

Your €uromole contributor misses some very important points, and ones that affect all code.

1. Elementary in programming is something called field validation. It's done by parsers that check a particular datum for specific attributes, one of which is size. If it's too big, don't accept it. If it's unusually or artificially huge (as an overstuffed buffer can be) then bail—if necessary by crashing yourself with a polling routine.

2. Some languages allow direct transfer to stacks or arrays. It's nice, and picks up speed in execution. It's also a recipe for disaster as the array can be stuffed inadvertently (see #1) with data that's out of size or type bounds. It's a good idea to have a similar parser that checks and validates arrays at various steps. Think of this as though it's the airline's second search of you before you get on the plane to Tel Aviv.

3. Windows was designed and coded in an older era. Cracks weren't thought of as a nuisance back then, because there was virtually no Internet connectivity prior to 1995; standards, therefore, were different. This isn't to apologize for bad coding, but the dependencies built on the Win32 API sets (and COM/COM+/DCOM etc) have resulted in huge problems in this rusting hulk of a kernel.

Hope all is well.

Tom Henderson

Hello,

I wonder how the likes of AS/400, MPEix, VMS, etc. would stand up if it had to stand the kind of scrunity Windows gets for being the most common platform around.

Even Linux users are starting to find increased vulnerabilities as the number of desktop increase. Nothing is secure.

However somewhere along the line we accepted the fact that if we use software, that bugs are enviable. Imagine a TV set that doesn't show yellow on every third Friday. It would be returned in the shop. If the retailer said you can only return it in its original packaging unopened then there are laws to stop this practice.

Once we let all the software vendors know that we won't pay for buggy software, then things will get better. I think.

EM

Mike

Nice to see someone giving the old DEC OSes kudos...

The PDP-11 processors, especially the later ones, like the PDP-11/70, -11/44, and later, had hardware memory protection, in the form of separate memory mapping registers for Kernel (K), Supervisor (S) and User (D), each of which were further subdivided into Instruction (I) space and Data (D) space. This was implemented primarily to get around the small address space of 4 megabytes, but also provided memory protection. RSX-11 used this feature as did other of the PDP-11 OSes.

VMS (or OpenVMS, if you prefer), extended this concept as you noted by the use of PSECTs, with the default protection on a PSECT being code, read-only. The processor naturally had the protection logic built-in as part of the virtual memory mapping hardware.

I agree that Microsoft should be brought to bear by one or several political entities for it's abusive attitude towards users. I wonder if Elliot Spitzer (Atty Gen'l here in NY) would be game. He's coming off a huge win over the Wall Street firms. I could see him taking this on, especially if he were to pursue it for the losses it has cost the state government, not to mention the average slob on the street.

What really cranks me most is that, as you pointed out, Microsoft has had access to the technology and the people to do it right and has squandered it away. Of course, from what I understand of their maverick culture, I'm not surprised. After all, if you are a multi-millionaire programmer, who is lionized, why would you want to listen to a couple of corporate weenie ex-DEC types? Shakes head...

JFW

Let me preface this rant with the fact that I have always enjoyed the commentary of the INQ, but I think you guys were WAY off base with the reachability of websites story. Have you given any consideration to the thought of 50 million folks in the US losing internet access almost simultaneously for an average of over twelve hours? Most websites of any magnitude have battery backup and generators to keep them humming, so there should have been minimal drop off of sites with a significant drop in users attempting to reach them.

Brian King

You wrote:

"But some end users and channel players would like Microsoft to do something else, and supply them with a trade-in for their original XP install CDs. That would avoid similar problems happening if - unlikely as it seems - people have to re-install Windows XP from scratch or buy new accessories and the like."

Well... for people with a CD-RW drive, "rolling your own" updated install CD is always a possibility, as described here.

Not all that difficult, and AFAIK perfectly legal as long as you already own a valid Windows license.

I've used the described procedure successfully, to update my own Win2K SP2 install CD to SP4.

Name supplied

Regarding your article on a possible new mini-SP from Microsoft and people wanting their original CDs replaced...

They can actually do it themselves....

The existing Windows XP SP1/SP1A updates have the ability to "slipstream" back into the original Windows XP Install CD.

You can manually extract the SP1 patch to a directory, copy the WinXP CD to your hard drive, run the SP1 patch with the appropriate command line options to tell it to integrate into the XP CD install, then burn it all back to a bootable CD and, voila, new fully installable SP1/SP1A version of XP based on your original, legally licensed CD.

So, if MS releases a "SP1B" or somesuch with all the current patches built in (namely the DCOM RPC fix) then it should be just as easy to do the same and build your own SP1B image for when you have to do a clean re-install.

Basic directions for SP1

Unpack the SP1 update (full downloadable) to a directory, say "C:SP1".

Copy XP install CD to hard drive, say "C:WINXP"

Run: "C:SP1Updateupdate.exe -s:C:WINXP"

Wait til it completes...

You will also need the boot image off your original XP CD.. ISOBuster or similar works quite nicely for extracting it. I think you may be able to via Nero as well. It's about a 4K file.

Then, from Nero, or similar pick BOOT CD, Expert Settings, No Emulation, select the extracted boot image, a load segment of 07C0 and 4 sectors.

Then select the contents of your WINXP directory and burn away....

A little tough for the novice perhaps but if someone whose main job is IT and is responsible for maintaining a corporate network and doing system installs can't handle it, then they should probably be replaced by someone competent anyway!

Michael Burkey

Radioactive Wasps

Mike we are used to such things out West.

Heck; The wasp problem is just another glowing review of Hanford.

Besides don't you know we grow thing BIGGER out West anyway.

Any day now expect a story on how Texas has the same problem and expects bigger insects. Besides the Wasps are how we control the Big Spiders (see 1950's Movie).

Steve Dawson

Madeira and Portugal

Madeira is not a "former Portuguese colony", because it isn't independent! It still is very much part of Portugal, just like the Azores.

They are both so called "autonomous regions", have autonomous governments and so on, but still belong to Portugal. You won't see Madeira's President easily admit it, but the islands (yes there are two, Madeira and Porto Santo) are actually quite dependent on money sent from continental Portugal. He keeps winning elections over there promising more and more independence from the continent, while ignoring this "minor" fact.

Yes, I am Portuguese and a fan of The Inquirer too. Just wanted to help you get your facts straight. Keep up the great work.

Regards,

Nelson Cruz

PC Manias

The Infernet and the Interweb

Mike,

What's with this "interweb" word? It sounds like something made up just for the sake of the author being able to say "I made that up". Is "web" not sufficient? Do you know its origins?

Cheers

Mark Smith