wu :: forums - Print Page


    
      
        wu :: forums
        (http://www.ocf.berkeley.edu/~wwu/cgi-bin/yabb/YaBB.cgi)
      

        riddles >> general problem-solving / chatting / whatever >> The Joel Test
        
(Message started by: BNC on Feb 13^th, 2006, 1:11pm)

Title: The Joel Test
Post by BNC on Feb 13^th, 2006, 1:11pm

Was this site (http://www.joelonsoftware.com/articles/fog0000000043.html) posted before?

If not, what do you think?

Title: Re: The Joel Test
Post by Icarus on Feb 13^th, 2006, 5:14pm

Not being a programmer, I can't really say much about it, but this does stand out:

Quote:

The truth is that most software organizations are running with a score of 2 or 3, and they need serious help, because companies like Microsoft run at 12 full-time.

Considering the quality of Microsoft product, I would say the test is far worse than he indicates. Then again, since one of the tests is:

Quote:

5. Do you fix bugs before writing new code?

Microsoft surely can't score better than 11.

(No, I am not an anti-MS nut. But I did buy Microsoft Millenium Edition once, and I may never forgive them for it!)

Title: Re: The Joel Test
Post by BNC on Feb 13^th, 2006, 9:27pm

I think the idea is that they fix knownbugs before writing new code -- whuch doesn't mean they don't release the sofware with known bugs...

(And, yes, I'm not much of a M$ fan either)

Title: Re: The Joel Test
Post by towr on Feb 14^th, 2006, 12:48am

I think amichail posted a few links to Joel's pages a while ago.
The article about the hungarian notation (http://www.joelonsoftware.com/articles/Wrong.html) is also interesting (basicly it means that in naming a variable you include some relevant information as to how it is to be used.)

Title: Re: The Joel Test
Post by Icarus on Feb 14^th, 2006, 3:37pm

That is interesting. When I saw recommendations that all variable names be prefaced with their type, I also rebelled. But he has done an excellent job of explaining what the real intent was, and the real notation makes a lot of sense.

(And I should think that fixing known bugs before subjecting your customers to them would be even wiser than doing it before writing new code, but apparently M$ execs don't agree.)

Title: Re: The Joel Test
Post by Grimbal on Feb 16^th, 2006, 8:40am

But the classification of strings as "unsafe" and "safe" is not the right thinking. He should rather distinguish between strings and HTML fragments. A string is a sequence of symbols to be displayed as such. An HTML fragment is something that includes HTML tags and escapes and is suitable for inclusion in an HTML page.

In this way, one could write:

sUser = (whatever the user entered)
htmlUser = htmlEncode(sUser)
htmlGreeting = "Hello, <b>" + htmlUser + "</b>";
write(htmlGreeting + "<br>");

Title: Re: The Joel Test
Post by Grimbal on Feb 16^th, 2006, 8:53am

on 02/14/06 at 15:37:09, Icarus wrote:

(And I should think that fixing known bugs before subjecting your customers to them would be even wiser than doing it before writing new code, but apparently M$ execs don't agree.)

M$ isn't known to produce great software. It's known to make money out of it. They know perfectly well that with an unstable program people will "do something about it" and purchase the next, hopefully more stable, version of the program.

Should they fix the bugs (for free) and discourage people from upgrading or should they put the effort in the next version that they can sell?

Title: Re: The Joel Test
Post by Icarus on Feb 16^th, 2006, 3:26pm

A short-sighted business practice, at best. You can't suppress the competition forever, and the bugs give your company a bad image. Eventually, it will return to bite you.

American automotive companies ruled the U.S. market in the 1950s and 1960s. They used their power to suppress competition and ignored quality issues. Eventually, the clear superiority of foreign models was so apparent that people began buying them despite all the protections and appeals to patriotism. The result is that every major U.S. auto manufacturer of the time is either gone, or has undergone major financial set-backs. They are still playing catch-up when it comes to quality issues, and suffering from the bad image they allowed to develop.

It may take awhile, but in the future M$ execs will rue the policies of M$ now.

Title: Re: The Joel Test
Post by Grimbal on Feb 17^th, 2006, 5:12am

BTW I found something missing in The Joel Test: automated tests and non-regression testing.

All too often, when adding a feature or fixing a bug, it breaks something else or reveals a bug that was hidden before. You need to run a comprehensive test suite on every daily build.