Petreley writes:
The only conclusion I can draw is that, like most commercial software developers, many open source
programmers are just plain lazy about proper error handling. But we're supposed to be better than that...
Doing good error / exception handling is tedious, requires hard thought, and is nowhere near as fun as writing the mainline code. It requires a lot of discipline to first sit down and think through an exception-handling policy, to stick with it while coding, and then to audit the code and ensure that everything sticks to it.
Lots of commercial projects, in the rush to get out the door, skip all the above, or some. Open source projects may or may not have the philosophy worked out, let alone the discipline.
I guess we need a small army of exception-handling drill sergeants looking over code and yelling "YOU CALL THAT ERROR HANDLING? MY GRANNY DOES BETTER ERROR HANDLING THAN THAT..."
horribly truncated sig:
When are Newsgroup Filters coming to Mozilla?
on
Mozilla 0.9.5
·
· Score: 1
I'll happily use Mozilla for everything once it has newsgroup filters just as it has filters for mail.
Any of you splendid Mozillons know if that's on the roadmap?
Seems to me the important thing about the "Ports" system is that it automatically fetches the system and all its dependencies and compiles them in one go.
THe most annoying thing about dealing with RPMs is that I don't find out until "rpm -i" time that it needs 6 other RPMs.
I have worked with PG (well, worked at the same company as he) at the late, little-lamented Lisp Machine maker Symbolics, and had some interaction with him concerining another company or two he's been at after Symbolics.
From my experience I can say the following: There seems to be a record of Phil winding up in litigation with companies he's been involved with. He has a certain arrogance, but you need that to survive the MIT environment; so did many of my Symbolics colleagues, it's a selection thing.
I read the following conclusions from this tale:
1. Never allow the VCs in without stacking the board in your favor FIRST. Make sure it's up to the insiders' initiative to do so, and get it done.
2. If you want somebody with the same corporate culture (e.g., in this case, not 9-to-5 unless they're the ones at night), make damn well sure they've been working in it and have that value system before you engage him.
3. Don't whine about the thickness of the contracts; read and understand them and think hard about their implications. It is your Arse (Digita) that will be hosed by them.
4. If you lie down with the suits you will get up with the lint. The suits will protect each other; don't ever doubt that for a minute. If they think you're attacking them, they'll use all the resources at their disposal to protect each other, BECAUSE THEY UNDERSTAND EACH OTHER AND NOT YOU. If the VCs don't seem to connect with your model for the business, find other ones and/or carry on as before.
5. Either this court case will work for you or not. If it does, fine, pick up the pieces and continue. If not, PUT IT THE HELL BEHIND YOU AND MOVE ON.
Slashdot Mirror
Petreley writes:
The only conclusion I can draw is that, like most commercial software developers, many open source
programmers are just plain lazy about proper error handling. But we're supposed to be better than that...
Doing good error / exception handling is tedious, requires hard thought, and is nowhere near as fun as writing the mainline code. It requires a lot of discipline to first sit down and think through an exception-handling policy, to stick with it while coding, and then to audit the code and ensure that everything sticks to it.
Lots of commercial projects, in the rush to get out the door, skip all the above, or some. Open source projects may or may not have the philosophy worked out, let alone the discipline.
I guess we need a small army of exception-handling drill sergeants looking over code and yelling "YOU CALL THAT ERROR HANDLING? MY GRANNY DOES BETTER ERROR HANDLING THAN THAT..."
horribly truncated sig:
I'll happily use Mozilla for everything once it has newsgroup filters just as it has filters for mail.
Any of you splendid Mozillons know if that's on the roadmap?
Oh, Shudder! It makes me think of Fidel Castro...
Seems to me the important thing about the "Ports" system is that it automatically fetches the system and all its dependencies and compiles them in one go.
THe most annoying thing about dealing with RPMs is that I don't find out until "rpm -i" time that it needs 6 other RPMs.
Thanks to the Port porters...
From my experience I can say the following: There seems to be a record of Phil winding up in litigation with companies he's been involved with. He has a certain arrogance, but you need that to survive the MIT environment; so did many of my Symbolics colleagues, it's a selection thing.
I read the following conclusions from this tale:
1. Never allow the VCs in without stacking the board in your favor FIRST. Make sure it's up to the insiders' initiative to do so, and get it done.
2. If you want somebody with the same corporate culture (e.g., in this case, not 9-to-5 unless they're the ones at night), make damn well sure they've been working in it and have that value system before you engage him.
3. Don't whine about the thickness of the contracts; read and understand them and think hard about their implications. It is your Arse (Digita) that will be hosed by them.
4. If you lie down with the suits you will get up with the lint. The suits will protect each other; don't ever doubt that for a minute. If they think you're attacking them, they'll use all the resources at their disposal to protect each other, BECAUSE THEY UNDERSTAND EACH OTHER AND NOT YOU. If the VCs don't seem to connect with your model for the business, find other ones and/or carry on as before.
5. Either this court case will work for you or not. If it does, fine, pick up the pieces and continue. If not, PUT IT THE HELL BEHIND YOU AND MOVE ON.
The diatribe has (been) spoken.
Weav
Just another damned employee