[geeks] Dual Core Rules: your bugs will run twice as fast

[Mark]

On 13 Feb 2007, at 19:09, Charles Shannon Hendrix wrote:

>> Oddly enough in the week of using Safari on my Mac Pro I've have no
>> issues with the Universal Binary, however I always used to get
>> lockups/slowdowns/crashes on the same version of the PowerPC binary
>> on my iMac G4.<snip>
>
> A program should never crash because the CPU is slow.
>
> That makes absolutely no sense at all.

I didn't specifically state that, I said I get more issues on a  
slower machine. The crashes may be just the usual crashes you get  
using Safari (they exist - I said as much), just observations of  
general Safari usability. Most of the issues I had on the iMac, if  
we're gonna get specific, were lockups (spinning pizza wheels) that  
never unlocked. I have to say these have occasionally crashed the  
browser eventually. You're right, it *shouldn't* be the case but  
don't forget you have 2 different CPU architectures here, and from  
the Get Info, 2 different versions of Safari (PPC is PowerPC only -  
Intel is Universal).

Still, it was just an observation. I've only been using the Mac Pro  
for a week, and have had zero issues with Safari, but if Leopard  
isn't due out any time soon I guess it gives me plenty of time to  
experience some! :o)

>> Personally I think web browsers get a lot of flack for the inability
>> of may web designers (myself included) to code stuff right.
>
> If the browser is written correctly, how could what a web designer  
> does cause
> it to crash?
>
> Granted, WWW design is mostly crap, but that's no excuse for the  
> browser to crash.

There are 2 unassailable facts here, that it's a very fuzzy picture  
whatever angle you look from and time is finite in any development  
cycle. No one browser could possibly cope with every single little  
foul-up because the nature of the beast is the combinations are  
extremely vast. I find enough variations just between stuff I write  
now and stuff I wrote 6 months ago, never mind between one code set  
and another. Imagine that multiplied by every web developer in the  
world, multiplied by every browser development team, and also  
factoring in varying levels of competancy and adherence to different  
guidelines - it's a mess alright.

>> have issues, Firefox has it's memory leaks, Safari crashes from time
>> to time, Opera is just... Opera... The stem of the problems though is
>> the lack of understandable guidelines and set-in-stone rules for this
>> stuff. It makes web development, and subsequently also writing a
>> browser, a flaming nightmare. Standards are great, ain't they??
>
> Again, that makes no sense.  I agree that the standards are all a  
> mess, but
> that's still no reason for the program to crash.

With the level of diversity in web design something will inevitably  
come along that crashes your browser. I have never in all my years  
(15 of them at least) on the internet met a browser that didn't throw  
it's arms up in disgust occasionally and fall on it's ass at  
something on a web page somewhere. Sometimes it's not even  
reproducible. It's weird but it's a fact that I've only ever seen in  
browsers.

> Over the years I've been given a lot of bad specifications for  
> software I've
> written.
>
> But every time my code crashed, it was my fault, not the specs fault.

Yeh, and people take the same attitude with browsers, it's the  
programmers fault because they didn't take infinity+1 days to iron  
out every last glitch, and when they try to claim time restrictions  
meant they couldn't everyone thinks they are making excuses. It's the  
way the world works I guess.

> I might have been foolish to try and meet a bad spec, but  
> ultimately when my
> program dumped core, it was something I did wrong.

In all those cases you knew what you were working with. You had a  
spec. Browser developers are not in that enviable situation.  
Theoretically they have a spec, but if you were to write it down...  
well I wouldn't want to move the filing cabinets it was in lets put  
it that way.

-- 
Mark Benson

My Blog:
<http://mdblog.68kmac.org>
68kMac.org:
<http://www.68kmac.org>
Visit my Homepage: <http://homepage.mac.com/markbenson>

"Never send a human to do a machine's job..."