>> No. But you make it sound as if all M$'s problems are because computer 
>> hardware is unreliable.
> 
> IME, every time I am called because of a friend/family computer "not 
> working", it is nothing to do with MS.  I can give you a long list of 
> reasons, mostly due to other software, hardware malfunction or drivers, 
> but none MS.

Most of the problems I get to look at aren't due to M$ either - but a 
significant number of them are.

>> This is manifestly not the case. M$'s problems are because they 
>> produce poor quality products.
> 
> I use MS products daily on 3 or 4 machines and cannot remember the last 
> time one of them crashed or when MS Office behaved badly.

Actual OS crashes are fairly rare, assuming you use your computer in a 
sane mannar.

But I'm not just talking about complete crashes of the entire OS, or 
even crashes of a single application. I'm talking about the whole 
quality equation - how M$ products in general tend to be unecessarily 
complicated, poorly documented, resource-inefficient, insecure, and so 
forth.

> You mention corrupted Word documents, but IIRC we decided that was due 
> to everyone using the same template/file that was corrupted and 
> spreading this corruption through all your documents.  This is not 
> really a fault of MS Office, especially when you didn't even use the 
> "Open and Repair" command which would have probably fixed the problem.

Let's suppose that a particular Word document is corrupted. Why should 
that make Word crash? Shouldn't it just pop up a message saying "I can't 
read this file, it seems to be corrupted"? Isn't that what "graceful 
failure" is all about? But no, Word just crashes outright.

I opened the same file in OpenOffice, and it just opened up as if there 
was nothing wrong with it. I saved it again, and it has worked in Word 
ever since.

Why is it that Word, a premium product designed and produced by the 
richest software company on earth, cannot do something that OpenOffice 
can? The people developing OO didn't even have access to a description 
of the file format; they had to reverse-engineer it. And yet, they 
somehow did a better job than the people who *designed* that file 
format. How can that be right??

(Let us not go into the fact that Word costs almost infinity times more 
than OpenOffice to start with...)

Post a reply to this message

> how M$ products in general tend to be unecessarily complicated, poorly 
> documented, resource-inefficient, insecure, and so forth.

TBH I've found MS products to be really well documents, in fact I would say 
better than any other software I've used.  Really, even if you have a 
complicated thing you want to do in Excel, the documentation usually has the 
answer.

> Let's suppose that a particular Word document is corrupted. Why should 
> that make Word crash? Shouldn't it just pop up a message saying "I can't 
> read this file, it seems to be corrupted"? Isn't that what "graceful 
> failure" is all about? But no, Word just crashes outright.

Yeh, they should just use the code from "Open and repair" for the normal 
"Open" operation, and if there were no faults just act silently.  OTOH maybe 
load times would increase significantly for large documents?

> Why is it that Word, a premium product designed and produced by the 
> richest software company on earth, cannot do something that OpenOffice 
> can? The people developing OO didn't even have access to a description of 
> the file format; they had to reverse-engineer it. And yet, they somehow 
> did a better job than the people who *designed* that file format. How can 
> that be right??
>
> (Let us not go into the fact that Word costs almost infinity times more 
> than OpenOffice to start with...)

I think you've answered your own questions there anyway, MS has to make 
money so they have all sorts of constraints that OpenOffice doesn't.  If an 
OpenOffice update is delayed by 6 months because they are fixing the 
loading-corrupt-files code, nobody can complain. But if MS attempts to delay 
Office 2013GT by 6 months because they want to fix the loading-corrupt-files 
code, they will likely be forced to release it anyway by the financial 
people.

Post a reply to this message

>> how M$ products in general tend to be unecessarily complicated, poorly 
>> documented, resource-inefficient, insecure, and so forth.
> 
> TBH I've found MS products to be really well documents, in fact I would 
> say better than any other software I've used.  Really, even if you have 
> a complicated thing you want to do in Excel, the documentation usually 
> has the answer.

Well, there are plenty of open-source projects that have *no* 
documentation at all. OTOH, most of what M$ provides leaves much to be 
desired. (E.g., I managed to find a KB article explaining how to use the 
Recovery Console. Except that, actually, it just duplicates, word for 
word, the terse command help built into the Recovery Console...)

In general, if you're trying to do something simple (e.g., how do I 
change the IE start page?), the documentation tells you. If you're 
trying to do anything moderately nontrivial, the documentation tends to 
not help at all... I guess this says something about their intended 
target audience?

>> Let's suppose that a particular Word document is corrupted. Why should 
>> that make Word crash? Shouldn't it just pop up a message saying "I 
>> can't read this file, it seems to be corrupted"? Isn't that what 
>> "graceful failure" is all about? But no, Word just crashes outright.
> 
> Yeh, they should just use the code from "Open and repair" for the normal 
> "Open" operation, and if there were no faults just act silently.  OTOH 
> maybe load times would increase significantly for large documents?

I don't know, but you'd think they could at least install an exception 
handler around the load routine so that if it fails, it doesn't crash 
all of Word, just the document load engine. Or *something*...

>> Why is it that Word, a premium product designed and produced by the 
>> richest software company on earth, cannot do something that OpenOffice 
>> can? The people developing OO didn't even have access to a description 
>> of the file format; they had to reverse-engineer it. And yet, they 
>> somehow did a better job than the people who *designed* that file 
>> format. How can that be right??
>>
>> (Let us not go into the fact that Word costs almost infinity times 
>> more than OpenOffice to start with...)
> 
> I think you've answered your own questions there anyway, MS has to make 
> money so they have all sorts of constraints that OpenOffice doesn't.  If 
> an OpenOffice update is delayed by 6 months because they are fixing the 
> loading-corrupt-files code, nobody can complain. But if MS attempts to 
> delay Office 2013GT by 6 months because they want to fix the 
> loading-corrupt-files code, they will likely be forced to release it 
> anyway by the financial people.

OTOH, OO is developed by a bunch of bored boffins in their spare time, 
whereas M$ can afford to hire the brightest people in the business and 
pay them to work on the problem 9-5 every single day. Given the hugely 
superior resources available to M$, you'd *think* they could produce a 
product that works properly under forceeable circumstances...

Post a reply to this message

> In general, if you're trying to do something simple (e.g., how do I change 
> the IE start page?), the documentation tells you. If you're trying to do 
> anything moderately nontrivial, the documentation tends to not help at 
> all... I guess this says something about their intended target audience?

Dunno, I've found lots of quite detailed articles before.

eg googling "windows recovery console" give me

http://support.microsoft.com/kb/314058

which tells me that you can use bootcfg without the /scan to avoid spending 
hours scanning all your drives.

> I don't know, but you'd think they could at least install an exception 
> handler around the load routine so that if it fails, it doesn't crash all 
> of Word, just the document load engine. Or *something*...

Maybe the load routine works fine, it just loads some corrupted data so when 
another part of the program comes to use it that bit crashes?

> OTOH, OO is developed by a bunch of bored boffins in their spare time, 
> whereas M$ can afford to hire the brightest people in the business and pay 
> them to work on the problem 9-5 every single day. Given the hugely 
> superior resources available to M$, you'd *think* they could produce a 
> product that works properly under forceeable circumstances...

I think you overestimate how many people MS has working on such things.  The 
main difference is MS can't just go and recruit 20 more people to work on 
something, unless it is going to be *profitable* for them to do so.  If the 
bored boffins at OO want to spend another 1000 man hours perfecting things, 
they can, they don't have to worry about profit.

Post a reply to this message

scott wrote:
>> In general, if you're trying to do something simple (e.g., how do I 
>> change the IE start page?), the documentation tells you. If you're 
>> trying to do anything moderately nontrivial, the documentation tends 
>> to not help at all... I guess this says something about their intended 
>> target audience?
> 
> Dunno, I've found lots of quite detailed articles before.
> 
> eg googling "windows recovery console" give me
> 
> http://support.microsoft.com/kb/314058
> 
> which tells me that you can use bootcfg without the /scan to avoid 
> spending hours scanning all your drives.

The command descriptions given here are a word-for-word reproduction of 
the interactive command help. It does not explain anywhere, for example, 
that bootcfg is merely a programatic way of editing BOOT.INI. (Also, it 
seems that /scan, /add and /rebuild *all* scan your disks for hours - 
I've tried it.)

*This* article explains it properly:

http://support.microsoft.com/kb/291980

>> I don't know, but you'd think they could at least install an exception 
>> handler around the load routine so that if it fails, it doesn't crash 
>> all of Word, just the document load engine. Or *something*...
> 
> Maybe the load routine works fine, it just loads some corrupted data so 
> when another part of the program comes to use it that bit crashes?

Thinking about it, the load routine probably does just dump a bunch of 
pointers directly into Word, which probably then trips the rendering 
engine over or something. (I've seen Word go into in infinite loop 
instead of crashing also. That, at least, you'd think would be 
detectable quite easily.)

>> OTOH, OO is developed by a bunch of bored boffins in their spare time, 
>> whereas M$ can afford to hire the brightest people in the business and 
>> pay them to work on the problem 9-5 every single day. Given the hugely 
>> superior resources available to M$, you'd *think* they could produce a 
>> product that works properly under forceeable circumstances...
> 
> I think you overestimate how many people MS has working on such things.  

Uh... they're the largest software corporation on the face of the Earth? 
(And the richest too.)

> The main difference is MS can't just go and recruit 20 more people to 
> work on something, unless it is going to be *profitable* for them to do 
> so.  If the bored boffins at OO want to spend another 1000 man hours 
> perfecting things, they can, they don't have to worry about profit.

Well, if M$ doesn't consider "making our products work properly" to be a 
"profitable" thing to do, then there's your answer. :-}

Post a reply to this message

>> I think you overestimate how many people MS has working on such things.
>
> Uh... they're the largest software corporation on the face of the Earth? 
> (And the richest too.)

But still they cannot hire or steer extra employees onto projects without 
justification.  That's the main point.  It's simply a matter of economics - 
if MS pushed more resources onto fixing every little bug they'd run out of 
money pretty quickly - it's just not feasible when you have to pay people to 
write/debug your code.

Post a reply to this message

scott wrote:
>>> I think you overestimate how many people MS has working on such things.
>>
>> Uh... they're the largest software corporation on the face of the 
>> Earth? (And the richest too.)
> 
> But still they cannot hire or steer extra employees onto projects 
> without justification.  That's the main point.  It's simply a matter of 
> economics - if MS pushed more resources onto fixing every little bug 
> they'd run out of money pretty quickly - it's just not feasible when you 
> have to pay people to write/debug your code.

Fixing every tiny bug might be infeasible, but you'd think they could at 
least fix the huge ones.

Anyway, somebody asked why people hate M$. And I gave an answer: because 
M$ does not consider performance or reliability to be "important" enough 
to make any attempt to improve them.

Fortunately, the makers of safety-critical hardware and software are 
bound by *laws* which prevent them from not fixing bugs just because 
it's "not profitable".

Post a reply to this message

> Fortunately, the makers of safety-critical hardware and software are bound 
> by *laws* which prevent them from not fixing bugs just because it's "not 
> profitable".

And as a result their software is orders of magnitudes more expensive, not 
exactly something you could sell to the masses...

Post a reply to this message

scott wrote:
>> Fortunately, the makers of safety-critical hardware and software are 
>> bound by *laws* which prevent them from not fixing bugs just because 
>> it's "not profitable".
> 
> And as a result their software is orders of magnitudes more expensive.

I'll grand you that.

OTOH, M$'s products aren't exactly "cheap", by any stretch of the 
imagination. I think what most people really object to is paying vast 
sums of money for a product that isn't actually all that good. Let's 
face if, when you run KLogic and it crashes for the 18th time, you think 
to yourself "oh well, it least it didn't cost me anything". If you paid 
£400 for a piece of software and it behaved the same way, you'd be 
pretty upset.

Post a reply to this message

> OTOH, M$'s products aren't exactly "cheap", by any stretch of the 
> imagination.

I don't think paying 2x-3x the price of a game for your OS and Office suite 
is expensive, and I guess most other people don't either.

> I think what most people really object to is paying vast sums of money for 
> a product that isn't actually all that good. Let's face if, when you run 
> KLogic and it crashes for the 18th time, you think to yourself "oh well, 
> it least it didn't cost me anything". If you paid £400 for a piece of 
> software and it behaved the same way, you'd be pretty upset.

Agreed, but my Windows or Office doesn't crash 18 times, and I didn't pay 
400 pounds for it :-)

However we did pay 10 thousand pounds for some CAD software, and that 
crashes occasionally (which is infinitly more often than the OS!).

Post a reply to this message