** UnhandledExceptionEventHandler :: OFFICIAL LC / CHATTER THREAD **

[kyleb]

lol, worst captchas of all time

http://www.docstoc.com/docs/1048763/...as-of-All-Time

some of these are ridiculously funny

[Neil S]

Some of those are terrible.

But I think the idea of using symbols and having those symbols correspond with letters is clever.

[Adam27X]

yeah I like that one and don't see why it's getting hate

[thunderchicken]

Originally Posted by diebitter

both imo. I think the ms sql server one would be particularly good for me, I know quite a few tricks about writing ****-off-a-shovel-fast sprocs.

That's my bread and butter.

[Neil S]

So yeah, I absolutely loathe MS Windows, and avoided having anything to do with that literally(!) headache-inducing platform for over a decade. But then I had a client willing to pay me a good, steady income, but needed me to have access to it.

I'll say one good thing for it: the platform desperately needed the addition it got of Windows Script Host. Don't know why it took 'em forever to add it, but there you go.

[_dave_]

Windows Script Host has been around for a looong time, certainly Win98, maybe 95. Or do you mean PowerShell?

[Neil S]

Originally Posted by _dave_

Windows Script Host has been around for a looong time, certainly Win98, maybe 95. Or do you mean PowerShell?

Don't recall it in 95 or NT 4, and that was when I escaped.

[_dave_]

Thinking back it was probably default from 98SE onwards, I think before that it came as part of the IE4 bundle that was well known for trashing systems with the lovely "Active Desktop".

I remember it being awesome, one time I had a job to collate daily sales figures going back years stored in Excel documents (one per day, all in random folders under "c:\accounts") lol. So I could either open up ~900 xls and find the numbers as was the plan, or... WSH wscript.exe to the rescue! did it in about 20 minutes (not counting the day it took me to learn / write the program). And I could be sure all the generated weekly/monthly/yearly totals were all accurate too easily scanning the generated big sheet that was a satisfying job imo!

[greg nice]

dont rewrite code

http://www.joelonsoftware.com/articl...000000069.html

[Adam27X]

Nice article.

When optimizing for speed, 1% of the work gets you 99% of the bang.

I'm skeptical about this statement though. Thoughts?

[jjshabado]

It's an exaggeration - but not a very big one (if you accept "work" to mean code changed).

[clowntable]

Quick design question:
I have a class for objects that I build by reading from a database i.e. I get all the attributes like name, width etc. from the database. Now the class will be extended to have an image attribute i.e. a .png file. Should I
1) save the image in the database as a BLOB and get it from the DB like all other properties
2) store it on the filesystem in say an images folder and store a reference to the location in the database i.e. image_path or something in the db is set to a string pointing to the location of the file
3) store it on the filesystem in say an images folder and get the location via code

I'm currently doing 3) i.e. if I had an object with a name attribute of 'hot_babe' I'd just store hot_babe.png in images and follow the convention of image = getImage('images/'+self.name+'.png') .. this is pseudocode, I use an operating system neutral way of actually finding the image

I feel like 1) would probably be better but BLOBs aren't the hottest stuff in the world either :P

[rt1]

Originally Posted by clowntable

Quick design question:
I have a class for objects that I build by reading from a database i.e. I get all the attributes like name, width etc. from the database. Now the class will be extended to have an image attribute i.e. a .png file. Should I
1) save the image in the database as a BLOB and get it from the DB like all other properties
2) store it on the filesystem in say an images folder and store a reference to the location in the database i.e. image_path or something in the db is set to a string pointing to the location of the file
3) store it on the filesystem in say an images folder and get the location via code

I'm currently doing 3) i.e. if I had an object with a name attribute of 'hot_babe' I'd just store hot_babe.png in images and follow the convention of image = getImage('images/'+self.name+'.png') .. this is pseudocode, I use an operating system neutral way of actually finding the image

I feel like 1) would probably be better but BLOBs aren't the hottest stuff in the world either :P

Absolutely #2. #2 and #3 are really the same thing. However, storing the relative location AND filename in the DB is going to provide more flexibility in the future without any addition cost right now.

Theres no point in storing the image in a blob column unless you need to run some db specific operation on the contents of the image file. #1 is also going to make reading those images quite costly.

[Neil S]

Originally Posted by greg nice

dont rewrite code

http://www.joelonsoftware.com/articl...000000069.html

I never liked this article. Joel's a smart guy, but I tend to find more use out of his posts about business models and strategy than his software engineering posts.

There are plenty of good reasons to rewrite code. Also to rewrite history, which is what he was going in this post anyway. Netscape was already on the way out by the time the Mozilla Project decided to start on Gecko and other rewritten components. To pretend they failed because of that is to ignore the fact that they were already doing so badly they had to run to the government to try to slap down the competitor that was just killing them, Microsoft, by offering a comparable product for free.

Who charges for a web browser these days? Only Opera, and even they give away free browsers where they're competitive (mobile).

[gaming_mouse]

Not a fan of the article either. I mean the overall point has its place, but there are plenty of good exceptions to the rule.