If people knew better how to open doors to the kinds of jobs they want, then your argument doesn't work. The reason why a college degree is needed is because everybody acts the same--like a herd of cattle. They send out resumes (moo!). Not to managers they know, or to managers who know them, but to some faceless HR dept (moooo). Of course a degree is handy then. How else to tell apart the Bovines?
So how to get a job then? How does RedHat hire folks? Do they put ads in local papers. No. They hire people they know. They don't scrummage through thousands of resumes looking for Linux kernel hackers. They turn to their employees, who say "Sure, that blizzard guy knows this stuff. We should hire him". The idea is that they hire people who they know can do the work.
Get good at what you do. Find companies that do what you do. Let them know who you are. Then a degree is just a piece of paper.
Job search sites are a poor way to find a job. Here's a recent quote from Ask The Headhunter over at the EE times:
It should be no surprise to you that the big career sites where you can post jobs and resumes are owned by advertising companies. The objective is not matching people with jobs; the objective is selling as many ads as possible. That's why none of these sites (that I know of) can or will report on hiring rates resulting from those postings.
If I'm wrong, and this site is different, then prove it. Show them stats!
When it was first released, I started building Mozilla on a P133 with 16M of memory. It built fine. With so little memory, you just have to make sure that you have plenty of swap space set up.
I'm running the Nav-only version of 4.5 on Linux right now and it's frustrating not being able to click on mail-to and nntp links. But I agree that the browser is most important. Maybe a better solution would be hooks for using your own mail and news clients (e.g. slrn and Mutt).
If that was implemented then you could divert work on the Mozilla Editor to the browser too. As I understand it, the Editor is needed right now to provide editing capabilities for News and Mail.
Someone else has already pointed out most of your mistakes. I'll add one more. Having the code available as tarballs via ftp doesn't do much good. The code changes too often, and then you get people complaining about old bugs, which adds more noise to the newsgroups. Besides, if you are interested in working on Mozilla, then updating using cvs saves you time too.
Bandwidth just wasn't the problem. Once you have pulled the SeaMonkey tree, updating your tree two or three times a week takes only about 15 minutes over a 33.6k modem.
Oops, screwed up a link on my last comment. I swear this one works.
I just hope that JWZ occasionally still posts to the mozilla newsgroups. Not that I read them just to catch his posts. Just like I don't read the newspaper just to read the comics.
What a journey: twentieth employee at Mosaic, then part of the fastest growing company ever, then an employee of the AOL juggernaut.
Read jwz's resignation if you haven't since it's a fascinating read. I found his explanation of why mozilla has failed to ship in a years time facinating. And I especially liked how he ended the paper by saying that mozilla's failures haven't been because of it's going open source, but mostly in spite of it (except he uses the word pixie).
It's easy to understand why he finally quit. Anybody who is there at the beginning, who works best when the pressures are highest and the excitement is greatest, is unlikely to fit in later on when a company grows large and boring.
I just hope he continues to write great code and the occasional gruntle prose and occasionally liven up the mozilla newsgroups.
The descriptions of the plane flights and talking to suits didn't give me the impression of a person enjoying it. And I also didn't get the impression he was looking for more of it. Where do you get that?
However, just like you, I do get the impression that he'd keep the job if conditions improved.
Which looks about as likely as pig flying, considering the number of people (or are they chimps at keyboards?) flaming him so far.
Remember all those/. articles you've read that had quotes from ESR? Their easy to remember because they almost always mention that Eric wrote "Cathedral and the Bazaar". Nearly everyone knows who wrote that paper.
And yet he didn't list being able to write as one of his qualifications. Even so he's already (26 comments so far), getting flamed for that list. Despite forgetting that you should be able to write too, he still gets flamed!
Fat chance finding anyone to replace him! It's hard enough to find a great coder, let alone one willing to work for free. And then ask that he be good at speaking in front of large audiences! Hehe. Suuuuuuure. No problem. Oh, and can he be creative and write really well too?
Don't think that he needs to be replaced? That's what I thought too. But we all know that just because you make a great product doesn't necessarily result in winning the game. And up until ESR retirement I'd just taken it for granted all those quotes and speaches. Now too late I understand the need. Hopefully enough of the people that flamed him before will too.
This is a good idea and I'm a piss poor writer, so I'm going to try again.
With a collaborative filtering scheme everybody gets to vote except AC's (for reasons that will become obvious). But you don't see comments scored by everybody--that would suck I'm sure. You only see scores by a group of people that score articles like you do.
The mighty Perl program that implements this has a big table it looks at. The columns are message ids for all the comments. The rows are users. The cells are scores.
What the program now does is find people who score like I do. That algorithm is the tough part to come up with. An example might be people that just score as close to me as possible. It might also use how often I score (scores per day/week) as a metric.
It results in better scoring, using more data, with everyones involvment, with an end to elitism and rules up the wazoo.
See the link I posted earlier to some good papers about all this. Like I said before--I can't write.
I'm a damn poor writer! With that in mind, realize that I missed collaborative filtering's best points in my comment. It's that you only see articles scored by people that vote like you. If you score "first posts" low, or not at all, then those comments get scored low by your "group".
I think you missed a point. With collaborative filtering comments are scored based on how others who are interested in the same kind of comments I'm interested in score. The only way that anti-linux comments would get filtered out, therefore, is if I also gave them low scores.
The idea is that you get a giant amount of data about comments. But rather than just clumping them all together you find people interested in the kind of stuff you like.
Your first point is a good one, because more data is a good thing for lots of reasons. But there is still a problem. All that accumulated data is just clumped together. If there was enough data (and/. is surely capable of that), then it would be better to get the opinions of people who liked the articles I liked. In the same spirit as people who like country music aren't likely to recommend an alternative CD that I'm likely to like.
The idea is you make all (logged in) users moderators. Everyone can vote. Gone are all the silly rules about giving away that you are a moderator. Gone is the elitism of a handful of moderators. Best, it'll work better, because there is more data available.
In effect, you will be choosing a group of people to act as personal moderators. They are choosen by selecting people with whom you have agreed with in the past.
This isn't my idea, and it isn't new. It was tried on a system called GroupLens, which was integrated into newsreaders like slrn, tin, and gnus. Check out the ps paper by Miller there.
The only downside is it might tax/. too much. Not sure. But I'm even willing to help if it's needed, because I think it's a Good Thing.
Operating Systems--Design and Implementation by Andrew S. Tanenbaum
Hardly used. Read only once, in fact. The are no markings in the book. Unlike most text books, you wont find highlighted sections with annotations marking all of the useful and helpful areas of the book.
Hey, I'd forgotten about that! See, now that I can respect. Randall (aka GWNICR) is obviously a thinking man and has some genuine concerns for the working spaceman.
I think that Eric Raymond has always gotten this one wrong too. If you reverse engineer software for the purpose of making keygens or to find out where to change a JZ to a JNZ, you are a cracker. The name makes sense.
If, OTOH you try to break into systems where your grubby hands don't belong, you are, some what unfortunately, a hacker.
They are two different things that need different names. Until another name is invented for a hacker, you'll just have to live with it. Hell, just call yourself a programmer.
Know why? Because shipping any game is tougher than shit. At the start, everything is exciting and the enthusiasm is high. Writing new code is fun. But at the end, when all you're doing is fixing bugs life is miserable, and your only goal is getting the f****er out the door.
Not that that is unique to writing games. But games make it tougher by putting a premium on performance (well, not SSI maybe;), and how quickly they become yesterdays technology.
So since it's not going to be used a decade later, there is no reason to sweat over making it look like the standard C library. Just point out to a kid looking at the code that here is an example of a program that runs on millions of machines without a hitch. I think you'll see some beauty in it then.
Probably more interesting would be to make a text front end for Mozilla. You could use the code from the old Mozilla that converted HTML to text to get started. I'm sure you could get a lot of help from the mozilla folks (try the netscape.public.mozilla.* newsgroups)
And use what? If you want an ergo keyboard, there isn't another good one around. I checked several computer stores (best buy, compusa, computer city) and couldn't find any other non-cheapo ergo keyboard. As for the Kenesis, they are too expensive. I paid $30 for my MS Natural (after $10 rebate), which is hard to beat. That Kenesis might be good, but not ten times as good!
Slashdot Mirror
If people knew better how to open doors to the kinds of jobs they want, then your argument doesn't work. The reason why a college degree is needed is because everybody acts the same--like a herd of cattle. They send out resumes (moo!). Not to managers they know, or to managers who know them, but to some faceless HR dept (moooo). Of course a degree is handy then. How else to tell apart the Bovines?
So how to get a job then? How does RedHat hire folks? Do they put ads in local papers. No. They hire people they know. They don't scrummage through thousands of resumes looking for Linux kernel hackers. They turn to their employees, who say "Sure, that blizzard guy knows this stuff. We should hire him". The idea is that they hire people who they know can do the work.
Get good at what you do. Find companies that do what you do. Let them know who you are. Then a degree is just a piece of paper.
Job search sites are a poor way to find a job. Here's a recent quote from Ask The Headhunter over at the EE times:
It should be no surprise to you that the big career sites where you can post jobs and resumes are owned by advertising companies. The objective is not matching people with jobs; the objective is selling as many ads as possible. That's why none of these sites (that I know of) can or will report on hiring rates resulting from those postings.
If I'm wrong, and this site is different, then prove it. Show them stats!
When it was first released, I started building Mozilla on a P133 with 16M of memory. It built fine. With so little memory, you just have to make sure that you have plenty of swap space set up.
I'm running the Nav-only version of 4.5 on Linux right now and it's frustrating not being able to click on mail-to and nntp links. But I agree that the browser is most important. Maybe a better solution would be hooks for using your own mail and news clients (e.g. slrn and Mutt).
If that was implemented then you could divert work on the Mozilla Editor to the browser too. As I understand it, the Editor is needed right now to provide editing capabilities for News and Mail.
Someone else has already pointed out most of your mistakes. I'll add one more. Having the code available as tarballs via ftp doesn't do much good. The code changes too often, and then you get people complaining about old bugs, which adds more noise to the newsgroups. Besides, if you are interested in working on Mozilla, then updating using cvs saves you time too.
Bandwidth just wasn't the problem. Once you have pulled the SeaMonkey tree, updating your tree two or three times a week takes only about 15 minutes over a 33.6k modem.
I'd guess not. His work on Mozilla.org was as evangalist, not as a programmer. Check Bonsai, but I don't think he's checked in code for a long time.
Oops, screwed up a link on my last comment. I swear this one works.
I just hope that JWZ occasionally still posts to the mozilla newsgroups. Not that I read them just to catch his posts. Just like I don't read the newspaper just to read the comics.
What a journey: twentieth employee at Mosaic, then part of the fastest growing company ever, then an employee of the AOL juggernaut.
Read jwz's resignation if you haven't since it's a fascinating read. I found his explanation of why mozilla has failed to ship in a years time facinating. And I especially liked how he ended the paper by saying that mozilla's failures haven't been because of it's going open source, but mostly in spite of it (except he uses the word pixie).
It's easy to understand why he finally quit. Anybody who is there at the beginning, who works best when the pressures are highest and the excitement is greatest, is unlikely to fit in later on when a company grows large and boring.
I just hope he continues to write great code and the occasional gruntle prose and occasionally liven up the mozilla newsgroups.
The descriptions of the plane flights and talking to suits didn't give me the impression of a person enjoying it. And I also didn't get the impression he was looking for more of it. Where do you get that?
However, just like you, I do get the impression that he'd keep the job if conditions improved.
Which looks about as likely as pig flying, considering the number of people (or are they chimps at keyboards?) flaming him so far.
Remember all those /. articles you've read that had quotes from ESR? Their easy to remember because they almost always mention that Eric wrote "Cathedral and the Bazaar". Nearly everyone knows who wrote that paper.
And yet he didn't list being able to write as one of his qualifications. Even so he's already (26 comments so far), getting flamed for that list. Despite forgetting that you should be able to write too, he still gets flamed!
Fat chance finding anyone to replace him! It's hard enough to find a great coder, let alone one willing to work for free. And then ask that he be good at speaking in front of large audiences! Hehe. Suuuuuuure. No problem. Oh, and can he be creative and write really well too?
Don't think that he needs to be replaced? That's what I thought too. But we all know that just because you make a great product doesn't necessarily result in winning the game. And up until ESR retirement I'd just taken it for granted all those quotes and speaches. Now too late I understand the need. Hopefully enough of the people that flamed him before will too.
That would be cool. Never thought of that.
This is a good idea and I'm a piss poor writer, so I'm going to try again.
With a collaborative filtering scheme everybody gets to vote except AC's (for reasons that will become obvious). But you don't see comments scored by everybody--that would suck I'm sure. You only see scores by a group of people that score articles like you do.
The mighty Perl program that implements this has a big table it looks at. The columns are message ids for all the comments. The rows are users. The cells are scores.
What the program now does is find people who score like I do. That algorithm is the tough part to come up with. An example might be people that just score as close to me as possible. It might also use how often I score (scores per day/week) as a metric.
It results in better scoring, using more data, with everyones involvment, with an end to elitism and rules up the wazoo.
See the link I posted earlier to some good papers about all this. Like I said before--I can't write.
I'm a damn poor writer! With that in mind, realize that I missed collaborative filtering's best points in my comment. It's that you only see articles scored by people that vote like you. If you score "first posts" low, or not at all, then those comments get scored low by your "group".
Arggh. Read the paper if you have the time.
I think you missed a point. With collaborative filtering comments are scored based on how others who are interested in the same kind of comments I'm interested in score. The only way that anti-linux comments would get filtered out, therefore, is if I also gave them low scores.
The idea is that you get a giant amount of data about comments. But rather than just clumping them all together you find people interested in the kind of stuff you like.
Your first point is a good one, because more data is a good thing for lots of reasons. But there is still a problem. All that accumulated data is just clumped together. If there was enough data (and /. is surely capable of that), then it would be better to get the opinions of people who liked the articles I liked. In the same spirit as people who like country music aren't likely to recommend an alternative CD that I'm likely to like.
Or something like that...
The idea is you make all (logged in) users moderators. Everyone can vote. Gone are all the silly rules about giving away that you are a moderator. Gone is the elitism of a handful of moderators. Best, it'll work better, because there is more data available.
In effect, you will be choosing a group of people to act as personal moderators. They are choosen by selecting people with whom you have agreed with in the past.
This isn't my idea, and it isn't new. It was tried on a system called GroupLens, which was integrated into newsreaders like slrn, tin, and gnus. Check out the ps paper by Miller there.
The only downside is it might tax /. too much. Not sure. But I'm even willing to help if it's needed, because I think it's a Good Thing.
Operating Systems--Design and Implementation by Andrew S. Tanenbaum
Hardly used. Read only once, in fact. The are no markings in the book. Unlike most text books, you wont find highlighted sections with annotations marking all of the useful and helpful areas of the book.
Hey, I'd forgotten about that! See, now that I can respect. Randall (aka GWNICR) is obviously a thinking man and has some genuine concerns for the working spaceman.
Ok, I just read the review of Empire Strikes Back. I can't believe he hated it! He gave a glowing review to Star Wars too.
Now, what I'm wondering is, is it still within the realm of common decency to send him a little email about what I think of his review?
I better read his review of Jedi first I guess...
I think that Eric Raymond has always gotten this one wrong too. If you reverse engineer software for the purpose of making keygens or to find out where to change a JZ to a JNZ, you are a cracker. The name makes sense.
If, OTOH you try to break into systems where your grubby hands don't belong, you are, some what unfortunately, a hacker.
They are two different things that need different names. Until another name is invented for a hacker, you'll just have to live with it. Hell, just call yourself a programmer.
Natalie Portman is the girl. Real good actor. Saw her in Beautiful Girls, The Professional, and some Woody Allen movie where everybody sang a lot.
So long as she doesn't break into song again, she should be great for the part.
Know why? Because shipping any game is tougher than shit. At the start, everything is exciting and the enthusiasm is high. Writing new code is fun. But at the end, when all you're doing is fixing bugs life is miserable, and your only goal is getting the f****er out the door.
Not that that is unique to writing games. But games make it tougher by putting a premium on performance (well, not SSI maybe;), and how quickly they become yesterdays technology.
So since it's not going to be used a decade later, there is no reason to sweat over making it look like the standard C library. Just point out to a kid looking at the code that here is an example of a program that runs on millions of machines without a hitch. I think you'll see some beauty in it then.
Probably more interesting would be to make a text front end for Mozilla. You could use the code from the old Mozilla that converted HTML to text to get started. I'm sure you could get a lot of help from the mozilla folks (try the netscape.public.mozilla.* newsgroups)
And use what? If you want an ergo keyboard, there isn't another good one around. I checked several computer stores (best buy, compusa, computer city) and couldn't find any other non-cheapo ergo keyboard. As for the Kenesis, they are too expensive. I paid $30 for my MS Natural (after $10 rebate), which is hard to beat. That Kenesis might be good, but not ten times as good!