Micrrosoft doesn't even fully understand their own PE format?

[Ben321]

Just was looking at this MSDN documentation on the PE format for EXE files.
https://msdn.microsoft.com/en-us/library/ms809762.aspx

Down in the description for the the Magic field (first field in the so called "optional" header), it says "Appears to be a signature WORD of some sort. Always appears to be set to 0x010B". Hey MS, don't know how your own executable format works? LOL fail Microsoft.

[Sitten Spynne]

That's not reference documentation. That's an article written by a person in 1994. The first three sentences of the article point out he isn't a Microsoft employee, but worked for a different company at the time.

So it's not a Microsoft article, and it's more than 20 years old. When he wrote the article, the PE format had just been changed a lot, and documentation for all the changes probably hadn't even been published. So he was having fun and guessing and wrote a magazine article, which is what we did before there were blogs.

Critical thinking failure.

[Ben321]

That's not reference documentation. That's an article written by a person in 1994. The first three sentences of the article point out he isn't a Microsoft employee, but worked for a different company at the time.

So it's not a Microsoft article, and it's more than 20 years old. When he wrote the article, the PE format had just been changed a lot, and documentation for all the changes probably hadn't even been published. So he was having fun and guessing and wrote a magazine article, which is what we did before there were blogs.

Critical thinking failure.

I didn't even see it wasn't an article. I assumed since it was on the msdn website that it was official documentation.

It was one of the first search results when I Googled for documentation on the PE header, so I assumed it was going to be official. I didn't think Google's search algorithm was that inaccurate.

[Sitten Spynne]

Learn real quickly algorithms aren't perfect. Google's sorts things based on, "Which page gets the most clicks?" because no algorithm can judge the content of the site. This tends to be very wonky for MSDN because:

• Older, out of date articles have lots of historic clicks.
• Usually when people get to MSDN and aren't on the right page, they start a new search from within MSDN.

For example, MSDN has fully moved to docs.microsoft.com, and MS is going to pull the plug on old MSDN one day. But almost all search results point to the VS 2010 version of documentation on MSDN still.

[TysonLPrice]

Hmmm...I'm detecting a pattern. Accuse Microsoft of not understanding their own PE format; nope - it wasn't them. Accuse Google of search algorithms that are inaccurate; nope - it wasn't them.

[jmcilhinney]

Hmmm...I'm detecting a pattern. Accuse Microsoft of not understanding their own PE format; nope - it wasn't them. Accuse Google of search algorithms that are inaccurate; nope - it wasn't them.

But why did this site let him post it if it wasn't true? You'd assume that a forum would have some sort of filter for that, right?

[Sitten Spynne]

False equivalence.

The original post is about attributing "Microsoft employee" status to the author of a magazine article posted on MSDN. The article clearly states, within the first 10 sentences, the author was employed by a third-party company. It's the equivalent of a 1996 hosted blog post, and requires some archaeology to decide if he wrote it "for fun" or "because the PE format changes weren't published yet but betas were available".

The next post was about how 'being the first Google result' isn't a measure of credibility. Google's algorithm, at a high level, is well-known. Many details of it are black-box, but it's always been true that top-status is a popularity contest. This is often exploited by "Google bombs" and as of yet the only mitigation Google seems to have is manual adjustments of the search results if they care about who it is ticking off.

Part of doing research is understanding how to analyze sources to determine credibility. If one believes "having the top Google result" is all that is needed, there are some very strange ideas one could be led to believe.