Hacker News Reader

Comment by kyberias | Hacker News Reader

This is a 11 year old knowledge base article for a bug in a product (Excel 97) that saw the light in 1997, that is 17 years ago. Yes, the second workaround is kind of hilarious, but let's not draw too many far-fetching conclusions from it. I believe many of us have seen crazier bugs.

View on HN

This is by far craziest I've seen: "Error Message: Your Password Must Be at Least 18770 Characters and Cannot Repeat Any of Your Previous 30689 Passwords" [1]

[1] https://support.microsoft.com/kb/276304

Edit: typo

"Note that the number of required characters changes from 17,145 to 18,770 with the installation of SP1."

Wow. The story behind the extra 1625 characters must be good.

My $0.02 is on a dull explanation: the max length and #remembered parameters must be configurable and/or depend on the access method, and the printf-like call that creates the message string must have passed erroneous pointers pointing into some DLL. When that DLL changed, or when another DLL changed in size, causing the DLL to move, the pointers pointed to different, but still constant data.

Whats crazy is Microsoft believing this is an acceptable solution rather than fixing/patching it.

They don't. It's listed as a workaround. The fix is Excel 2000.

The crazy part is they made you pay for the fix. Therefore, it was considered acceptable behavior for Excel 97.

Okay come on. It's not like this is a ridiculously common use case. Plus there are 2 other workarounds.

If you're playing with Oracle data that often, you're probably an enterprise customer who's going to buy the next version anyway.

It's not like this was happening every time you wanted to save or copy/paste data.

I agree it's not a very relevant problem that didn't affect too many people. Still, it shows a pretty lame bug (that is, if the Oracle folks did their job correctly). In 1997 it was already usual to offer updates through the internet.

It's still kind of embarrassing.

>I believe many of us have seen crazier bugs.

Might be, but is it a good sign for computer industry as a whole?

It's a sign we're having trouble in handling the complexity of our computer systems. Programming is mostly human-mind-bound now (and has been for many years); we need better tools and better abstractions.

I think we need to remove that complexity and find a simpler approach to things, not just "better" (more complex) tools to layer even more complexity on top of complex.

I also would recommend that approach! The problem is, that you need more professional and better programmers that are aware of the problem and do reduce the complexity. Many self indulgent programmers, that think they are smart, won't.

But the trend is more to hire more and more novice programmers (because they are cheaper) ... so we get more into trouble in the future!

Agree completely. I think that contemporary CS education is really responsible for this "more complexity is better" mentality, by encouraging overly general and abstracted solutions, and that abstraction should really be applied only when it's necessary and not as a "just because we can" thing.

Aspiring programmers could learn a lot from the demoscene, where astonishingly impressive things are done with very little code and complexity.

Let's rephrase my point: the existence of this knowledge base article should not be taken as a sign for ANYTHING. :)

>far-fetching conclusions

like bot detection on NT-based systems? Yes, lets not talk about that.

Excel does bot detection? :)

Reply on news.ycombinator.com