Looking at the links to kernel traffic posted elsewhere in this article, I see you're right about what Linus thinks about threads.
But as for state machines, in fact state machines were what I was writing all the times I have written my own scheduler, but I strongly suspect having a kernel developer write a thread scheduler once and have everyone use it is far more efficient than to have every application developer write their own state machines.
While the kernel can deal with it, user-space state machines can't deal with the problem of blocking on a resource. This can only be handled by a kernel scheduler. On a preemptive multitasking but single-threaded system, you achieve efficiency by scheduling a new process when a process blocks; one a multithreaded system, you can schedule a new thread, which may be any thread in any process (or to use the BeOS term, a "team").
There's no way to handle this with user-space threads of any kind, whether they're implemented as some kind of state machine or faked up to look like real preemptive threads.
To directly address the question, though, I think ethics are of the highest importance in the work of a consultant, and are probably the most important guide for you to follow, more important than writing good code. You at least have the hope of debugging bad code.
The question goes both ways though, clients are occassionally unethical and many clients who wouldn't think of screwing you if you were a full-time employee would be happy to short you for weeks of pay earned as a consultant.
You have to protect yourself, start early by finding a good attorney before taking on work - certainly before trouble starts - and have your attorney review all your contracts before you sign them.
Also trust in your feelings and don't do business with someone you feel is not ethical. It's just not worth the heartache.
The ZooLib cross-platform application framework has a platform independentent C++ thread and mutex library for Linux (and other POSIX), Win32, Mac OS (with handrolled threads), and BeOS.
I'm pretty perplexed why Linus should think that thread programming is harder. This makes me wonder whether he's ever really tried in a serious way - a lot of programs are much easier to write when you use threads.
Just ask anyone who's had to write their own task scheduler for private use within one program, as I have when writing my
Raindrops screensaver for the Mac, or when I wrote a test tool for MacTCP, that enables up to 64 TCP streams and UDP pseudo-streams between any combinations of a Mac on a network - I wrote that in 1990 at Apple, well before threads were available on the Mac OS and it was a pain.
But personally I don't see the point of having threads not implemented in the kernel. It means all your threads block if any of them do. That's the point of kernel threads. I think solaris threads work the way they do because the solaris schedular isn't so efficient, so it's more efficient to multitask some of your threads in user space.
This also relates to why multithreaded programs are more efficient and responsive than single threaded ones, even on uniprocessor machines. If your process blocks to await a resource in a single threaded program, you get to wait while time is wasted. This is OK for servers where avergage efficiency is the main concern but is not appropriate for interactive or near-real-time use.
With a multithreaded program, if one thread blocks for a resource, another thread can run, and if it's another thread in the same program your program's utilization of the CPU will be greater. This is especially the case if the kernel is preemptible and multithreaded, as it is on the BeOS and can be with experimental patches on the Linux kernel.
For an experiment, using the same machine, compare the interactice responsiveness of Windows, the BeOS and Linux, and I can assure you that even for uniprocessor machines the BeOS will be much more efficient, because all native BeOS programs are multithreaded. Threading is available on Linux and Windows but are not yet that widely used.
It's just a proposal so far, I want to find a few developers to collaborate with before I start, but I'm also writing and posting
articles there on the general topic of making Free Software of better quality and also testing the kernel in particular.
While this doesn't provide an answer to your question, some of what I plan would be useful to think about in selecting a bugbase.
The number one feature that I wished to have in bugbases at companies I've worked at in the past was for the user to have an ability to describe a configuration of a machine and then give it a name in the database.
Each user would have some number, one or many, preset machine configurations, particularly the hardware configurations in my case, and the components of these configurations would be drawn from a database describing every piece of hardware that could conceivably be placed in a Linux box.
Then when you go to report a bug, you log in (so your contact info and presets are available), select the config that has the bug, and describe it.
For the Linux kernel, you'd paste in or upload your.config file (this is created by the kernel configuration process).
Then kernel developers could do boolean searches, say "find me all the crashes involving TCP that had a WhizzyNet card installed but in which PPP is not configured in the kernel".
The other part of this is that it is meant to be easy to use for both those who test new kernels (to report bugs) and kernel developers (to research bug reports).
I got the idea to develop the database after subscribing to the linux-kernel mailing list to resolve a bug on my laptop during the 2.4.0-test series kernels.
I was able to work with the list to resolve the bug and see that the patch stuck in the kernel sources, but I felt that many people who might otherwise like to contribute to testing new kernels might be put off by the process of dealing with the list - it's not sufficient simply to report a bug, sometimes patches don't "stick" and you've got to hang around until you're sure your bug stays fixed, but linux-kernel has one of the highest amounts of traffic of any internet mailing list.
Tell that to the psychiatrists and psychologists who treated me in the psychiatric hospital where I spent the summer of 1985, when I was diagnosed with schizoaffective disorder.
Or to my insurance company at the time, who sprung for the $10,000 treatment.
I know it may be fashionable to claim one is manic depressive, but I can assure you it's unmistakeable in me, and it's not something I would choose. There's no Hell so deep as what I experienced in my college days.
If you don't believe me, maybe you'll at least believe I do some good in my writings. I regularly receive emails in response to my website on Manic Depression from people who've experienced it - either they have it or their loved ones do - and they don't know what to do about it.
There is a lot of information available about the illness but because of the stigma which is perpetuated by ignorant people such as yourself, not a lot of folks have the confidence in themselves to speak out about their experience publicly as I do, occasionally here on Slashdot and permanently on that website.
The people who write to me tell me the fact that I speak out personally as I do makes a tremendous difference in their lives. It has happened many, many times that someone has written to me to tell me that I'm the first person they've shared the fact that they think they're mentally ill, or contemplating suicide.
You should know that that page receives about 3,000 hits a month, and I recieve several emails a day from people looking for advice, about half of them people who have the illness and the others from people who are closely involved with a sufferer and don't know how to deal with them.
By the way, I'm happily married. I first met my wife three years ago, and we were married in St. John's Newfoundland on July 22, 2000. We just bought our first house together, in Midcoast Maine, and moved into it last weekend.
I was in and out of my University physics studies a number of times, and having a generally bad time, because of
a serious illness, and at some point decided I should get out and get a programming job because I figured I'd be better at that than school.
I didn't really know how to program, I knew a little FORTRAN, C and Basic from doing data analysis during summer jobs, and I didn't really like it all that much. I used to really have to struggle to spend several weeks writing a 500 line program, and I'm sure I'd be embarrassed if I had to look at the source code to those programs today.
I figured I'd program for a while because it paid the rent (I was making $20k a year doing Sun administration and writing image processing software), but when I figured out what I really wanted to do for a living I'd quit programming and get a real job.
That was in 1988. Then some consultant visited and installed GNU Emacs on our machines (two Sun 3/160's, one diskless, both with terminals and no workstation monitor, but with frame grabber cards and NTSC color monitors). He explained about the GNU manifesto.
I thought it was pretty cool but didn't see it affecting me personally in a big way. I was mostly annoyed that I had to wait up while the consultant installed the software on what was supposed to be my day off while a ladyfriend was visiting from away.
Then my friend Jeff Keller, who went to MIT for a while and vaguely knew Richard Stallman, spent an evening with me singing the praises of Emacs. What I really wanted was VI with macros you could program to include conditional branches, and he said it had all though and much much more.
So I learned to actually use Emacs, and soon learned that it was quite extensible, but it wasn't made too clear how to extend it. The online manual was useful mainly to people who already knew what they were doing.
So I read the source code. One thing I was interested in doing was writing C functions that were callable from Emacs lisp as lisp functions. There are many such functions built into Emacs (usually for performance) and you can add your own. There's this big DEFUN macro that even makes the C API look like Lisp.
I learned that and a lot more. I learned what an eloquent statement of software architecture Emacs is.
I learned that there really was something worth my while doing in the way of software.
I wanted to write a program like that someday. Not another big editor, but a program that would someday strike other young programmers the way Emacs struck me.
During the course of reading the source code, one day I stayed at my terminal 24 hours straight, arising only to get coffee and use the restroom, not even eating. I only realized how much time had passed when I started to fall asleep.
That was when I started to take programming seriously. I began to put serious effort into studying programming, and studying it deeply.
For example I would read Knuth's The Art of Computer Programming on the bus on the way to work and I would stay up all night after work learning to program better on my Macintosh at home.
For many years I selected all of my jobs based mainly on what I could learn from them.
So in a very direct and profound way I owe it all to Richard Stallman and Emacs.
I still haven't written my great program yet. I don't even know what it will be. One project I've worked on peripherally is the ZooLib cross-platform application framework and a project I've just started up but not gotten too far with yet is the Linux Quality Database.
I did finally get my B.A. in Physics, from UC Santa Cruz, but only after being out of school working at a programmer for a number of years.
Michael D. Crawford GoingWare Inc
The best mailing list there is for junk mail
on
Spammer Gets Spammed
·
· Score: 2
... of the paper, postal variety, is the Direct Marketing Association's Direct Mail Preference Service.
Yes, this is the list you can submit your name and address to indicate that you don't want to receive unsolicited commercial postal mail. And to some extent it will cut down on certain types of regular junk mail.
But my old boss at Working Software, Dave Johnson, who wrote the chapter on direct mail in
The High-Tech Marketing Companion, says that the Mail Preference Service has the very highest response rate of all - for certain kinds of product offers.
(For a long period of time Working Software made most of its sales through direct mail, and Dave became quite an expert on direct mail. This was after he nearly went bankrupt listening to "channel people".)
What kind of product offers sell through this list?
Studded dog collars, burglar alarms, personal security devices, gun magazines and in general products that are aimed at people who are concerned with personal security and just want to be left alone.
Being on the DMA opt-out list doesn't actually prevent you from receiving mail. Instead, members who care to bother (usually because they don't want to waste money sending mail to people who won't respond) get the list periodically and use it to prune their in-house lists. So for lists whose owners bother to go to the trouble, you will be taken off some lists.
But studded dog-collar vendors just take the list and print up mailing labels!
You'd be suprised what you can turn up on a search engine using keywords including "resume" plus company names.
Maybe the people M$ suspected were up to something didn't have their own resumes on the web, but suppose a low-level employee at their company did, and listed the ex-microsoft execs as references.
I have been told by clients who have actually signed with me to do paying work for them that they hired me because my resume on the web matched the name of some previous employer of mine they were searching for. In those cases, the clients figured that an engineer who worked at these previous companies would be qualified to do their work.
Got your own resume on the web? Does your hosting service give you full logs, not just hit count summaries? Grep out your resume URL and look at the referring pages - in many cases the referring page URL will include the search keywords someone used to find your resume. It's pretty interesting to see what people are entering and then clicking on my resume as a result.
When a company makes you a job offer, ask them to send you over any contracts they are going to ask you to sign. You could give them a fax number, have them express them, or come pick them up.
Don't have a fax? Maybe the local copy shop downtown will allow you to receive faxes for a fee. I've known drug stores where I received faxes regularly.
Read the contracts thoroughly. You want my best advice, no matter what it says, take it to an attorney.
Make it clear to your new employer that you're going to be paying to have an attorney review their contracts, and if they make any changes later you're going to have to consult an attorney again. If they give you a hard time about this let them know that you'll be billing them for the additional expense.
Now, it's very important that you've asked up front for your contract. If the contract hasn't arrived before the day you're to start work, call in and say you can't show up because you haven't had an opportunity to see the contract. Tell them, as earnestly as you can, that a contract is a serious commitment between you and them and you wouldn't dream of signing one without reading it first, at home, without them watching you, and having access to an attorney.
If they give you a contract you haven't seen before when you arrive, inform them that you'd made clear your desire to review any contracts ahead of time and walk out.
If they're really worth working for, they'll get you the contract to review. If you're the kind of person for which it would really be an issue to have a non-compete agreement at all (exec, high-level engineer etc.), then they're not just going to let you go work for someone else after you've made it clear they insulted you and tried to deprive you of your right to make informed decisions about your own future.
Did you get the job through a headhunter or contract employment agency? You should know that they're paying, at minimum, 30% of your first year's salary for the headhunter's fee to hire you, or 30% of your hourly rate to the agency for the entire time you work (not just the first year). Consider how valuable you must be if they're paying this much for you, so really it's not that much to ask that you be completely informed before showing up to work.
To make it crystal-clear: Your decision to accept the job or contract should be based in part on full information about the contracts you will be signing. Do not make this decision until you've read the contract. Tell them that acceptance of their offer letter is conditional on reading their contract first.
I'm not kidding about this and this is not a troll. I really mean this and I'm speaking from hard experience. I've been a software engineer for thirteen years now and have been running my own independent consulting business for almost three years (by independent, I mean I don't work through agencies, so I'm on my own when I sign a contract).
You think you really need a job? Tired of flipping burgers - or think that's what you'll be doing if you don't sign? Thimk of the months or years of misery that will result if you make a rash decision right from the start. Not only will you regret it, but you may lose tens of thousands of dollars as a direct result of taking this job, rather than turning it down and finding another, or negotiating a better contract.
There's something called "opportunity cost" that many managers know about but not a lot of regular employees. It is the cost of losing the opportunity to do something else. You may think you've gained by getting a job, but you may have lost far more by not turning it down and taking a different one, or getting rid of that noncompete so that when this job is done (and no high-tech job lasts forever) you can get the rewarding job of your choice.
The developers were out their shipping many fine, supported commercial products that worked very well - Gobe, Steinberg was onboard developing audio applications, Thomas Dolby spoke at the developer conference about supporting Be with his audio software (he was wildly enthusiastic).
No, the incompetence was not the developers. It was Be's press relations and marketing (or to be more precise, the lack thereof). Be's technical staff is highly competent, and they produced a fine product, something Linux developers would do well to study because of it's ease of installation, smooth and trouble-free performance, and clean integration. It's also a joy to program.
My wonderfully positive experiences with using the BeOS and my sadness at having to leave the platform behind in order to have currently supported hardware and software is another component of what motivates my effort to improve the quality of Linux and Free Software in general.
But Be's business management was always, and still is, incompetent. There was never any effort made to convince people to make the move to the BeOS - they were always waiting for more applications to appear, even when there were plenty of applications for the average user to do almost all their daily work.
Even though almost anyone who ever laid eyes on the system became an instant convert - and many of those became evangelists, such as I especially after I got it running on my laptop, Gassee never felt it was ready for the mass market and kept holding it back from the people who could have done him the most good - the public who was eager to try something better if only they knew it existed and where to get it.
For example, I never once saw a banner ad on a website that was run on a non-Be related site directing the user to Be's website or BeDepot (Be's now-defunct eCommerce site).
And speaking of BeDepot - Gassee very early on spoke repeatedly about the problem of penetrating software channels for new developers, especially developers for new platforms like the BeOS, and how Be was going to be very advanced in its thinking by offerring software for sale via download at BeDepot.
And we all have seen what has become of eCommerce in general - it's the only way you can get many products for mainstream platforms - but BeDepot was one of the most ineptly run operations I have ever seen.
BeDepot's ineptitude was murder to developers for whom it was the only sales channel - so you've got a newly released product? What do you do when it takes three months to appear on the website because they're either too understaffed to deal with it or too incompetent to just upload the damn file? I've heard lots of complaints from developers who had to wait months to receive their contract from Be that would enable them to sell on BeDepot.
It's not like it's so hard to run a good online Be software site, as demonstrated by the folks who founded BeBits, in part to make up for BeDepot's and BeWare's inadequacies. (BeWare was for free downloads).
If you won't listen to me because of Be's lack of character, listen to me because of Be's demonstrated incompetence, and know that if they couldn't keep an eCommerce site with a few dozen titles running efficiently, how are they going to provide adequate support for a bunch of disparate embedded hardware vendors?
By the way, you should note that many of Be's most prominent engineers left when Be made the decision to drop the desktop and move to Internet Appliances, folks like Dominic who wrote the BFS Journaled Filesystem (with indexed file attributes - there's a read-only version of BFS you can get for Linux, that I'd like to make read-write, and would make a great addition to the system).
Sony should listen, not because Be kicked me off their mailing list, but the reason I posted the criticism that got me kicked off:
After years of silver-tongued promises of support from Be for such things as marketing support for developers, Be never effectively marketed it's own BeOS desktop operating system and then announced that it was ceasing development of it, except to the extent necessary to support it's development of the BeIA internet appliance platform.
This means, for example, that the current rev of the BeOS won't install on my PC because I use an Ultra160 SCSI controller - Be doesn't see these as necessary for support of it's development platform.
And you can forget about marketing support for those of use who listened to Be's promises and actually shipped commercial products for the BeOS.
While I may have lost a lot of time and effort coding spellcheckers that could have been better spent elsewhere, Be developers BeatWare and Adamation lost millions, and are only able to survive because they ported to Windows or Mac.
My essential point is that Be is not to be trusted by Sony when times get difficult for it. There was a time when folks like me, Adamation and Beatware were all Be had. If Be can screw us, why can't they screw Sony?
If anyone from Sony is reading this, I'd like you to understand that Be, Inc. does not possess the moral fiber that is necessary for a company you will be trusting for a business-critical component of any product.
This lack of integrity goes to the heart of Be's corporate culture, and is in particular exemplified by the attitudes and practices of Be's CEO, Jean Louis Gassee.
Read about the experience of one of Be's formerly most-loyal and most enthusiastic business partners in:
Note: the page gives the URL's for a number of alternative vendors for Internet Appliance operating systems, many of which are open source (an important consideration in a market where your vendor may abandon you). Also see the Embedded Linux Journal.
The kind folks at SunSITE Denmark are helping me out with my proposal to make testing new Linux kernels more widespread and effective with the Linux Quality Database at:
So far it's just a proposal. In the short term there will be resources on testing strategies, as well as tips on writing quality code. I the long run will come a nice web form for reporting bugs in a way that will be especially meaningful for kernel developers (capturing and searching hardware configuration and kernel config options).
Even with a proprietary closed-source program, you at least still have the distribution media so you can still reinstall your application if the original publisher stops selling the program.
But with a web-based application, the publisher keeps the executable binary, and even with copylefted code, the GPL's source code distribution requirement does not come into play because the binary is never distributed.
This will ultimately lead to disaster for ordinary users as web applications make the move from news, shopping and entertainment to functions that people depend on in their daily lives or businesses.
When I was a QA engineer for MacTCP at A Big Fruit Company I knew where my knew cube was going to me before the poor bloke in it knew he was slated to be laid off.
My manager let it slip then asked me to keep quiet about it.
The reason for the bogus HTML above is that somewhere in the combination of Netscape 4.73 and XFree86 4.0.1 I've got a severe memory leak that will eventually result in my system paging so heavily it becomes unusuable.
And it always seems to happen when I'm composing something in a web form like here on Slashdot, so I end up losing it, so I immediately hit the submit button when it started happening while composing the above.
I have word from some folks who maintain mirrors of kernel.org that the mirrors are actually lightly loaded tonight. The problem is that www.kernel.org has the list of mirrors, and being slashdotted, no one can find a a mirror or know that an organized mirror system even exists.
If is very easy to figure out your nearest mirror:
Determine the two-letter country code for the nation you reside in. This will be the same as the top-level national domain, ca for Canada, uk for United Kingdom, ch for Switzerland and so on.
If you're using ftp, the server name starts with ftp. If you're using http (a web browser), then it starts with www.
Put the country code in the middle. Add.kernel.org at the end.
and so on. If you have a web browser, it is more efficient to access the servers via HTTP, as you don't use a process that sits idle during the time you're not downloading something as is the case with FTP.
The Linux kernels have their own numbering scheme which is independent of any distribution. And new kernels don't come out all that often. That's why, after all these years, the kernel is at version 2.4.0 tonight.
Distros update their versions for lots of different reasons, often because a number of new user programs are available, a new XFree86 version is available that is needed for new video cards, or sometimes for purely marketing reasons, as when Slackware jumped from 4.0 to 7.0 in order to reach marketing parity with Red Hat - Slackware generally used a conservative numbering scheme, but Red Hat advanced the numbers rapidly, even though all the distros are based on the same stuff, but lots of people gave the Slackware folks the message they didn't want software they perceived as out of date.
So no, it's fine to use the 2.4.0 kernel with Mandrake 7.1. The version of the kernel bears no explicit relation to the version of the distribution.
By the way, if you wonder where most of the rest of the programs on your distro come from, you have the fine folks at The Free Software Foundation to thank. Richard Stallman takes pains to point out the system is more properly called "GNU/Linux", because what Linus developed was a kernel that the already-existing GNU programs could run on.
If you're going to work with the new kernel, I highly recommand you browse the mailing list. But linux-kernel has one of the highest volumes of any internet mailing list so you probably don't want to actually subscribe (you don't have to subscribe to post, unlike many mailing lists).
Why settle for 2.4.0 when you can get the very latest?:
Linux 2.4.0ac1.
I got in before the Slashdot effect myself - I'm running 2.4.0-prerelease-ac5.
They have this time at Apple in the development of new hardware called freezing the ROMs. You have to do it so you can get them into production to include in the new machines. But development of ROM software doesn't stop, it just takes the form of RAM-based patches...
Well I guess I was hoping there would be more widespread testing before the -nothing release was released, because now that 2.4.0 final has been released, some distros are likely to use it right away rather than wait for bugs to stabilize and some people will put it straight into production use.
It's got some helpful, practical tips on downloading and building a kernel from sources as well as providing feedback to the kernel developers. All of the information in the article is available somewhere, but when I first began testing with 2.4.0-test1, I found some things difficult to figure out, so I felt that it would be nice to put what I learned all down in one place.
This is part of an overall effort to improve the quality of Free Software. Another part of the effort is the just-founded Linux Quality Database, so far just a proposal - contact me at crawford@goingware.com if you want to help.
And finally, some interesting trivia for you:
After a long beta testing period and many delays, Windows 2000 shipped with 64000 documented bugs of which 25000 were considered serious by Microsoft itself (the figures are quoted from memory, I might be a bit off). The Windows trade press reported that the opinion was widely held by IT managers that one should not install Windows 2000 on any machine until a few service packs had been released, with Windows 2000 server not being considered ready for use until much later than the desktop user version.
The BSD/Mach based Mac OS X, derived from the NeXT operating system and now in beta testing, is Apple's first operating system to support protected memory that is expected to be widely used. (Another was A/UX, Apple's Unix port, but it served only a niche market).
Apple has been trying for ten years, longer than Linux has been in existence, to write a modern operating system. The first was Pink, renamed Taligent when Apple collaborated with IBM on it. In the end all that came of it was a little-used object-oriented programming framework.
The next was Copland, and I don't know why it was never completed exactly, but I was offerred a job as a performance engineer on the Copland project when I worked at Apple, but I turned it down - I didn't tell them this but I had the sense that Copland was a project that would never ship.
And Linus wrote Linux when he was a college student, which combined with the GNU utilities forms the operating system we know and love today.
These huge, well-funded corporations can't get it together to program their way out of a wet paper bag but a bunch of freaks on the Net have written the fastest growing operating system in use today.
You think Aharonian's a hero? You don't know the half of it - he's being sued by an intellectual property holding company for patent infringement. The basis of their suit include such claims as the accusation that Aharonian
shamelessly, and oftentimes profanely, attack[ing] the United States Government, specifically the Patent and Trademark Office
The entries to the International Obfuscated C Code Contest frequently have the effect of printing out a text string that is not readily apparent from the source code.
In many cases you could change the text to say "Copyright (c) 2000 Free Software Foundation. This program is free software; yadda yadda".
If you're program is a GUI program, it may not be too hard to obscure the generation of this text and show it in a message box.
One thing I suggest though is that if you do this, you split up the code that does it and distribute it all over your program, and include several independent ways it can be done with independent implementations.
I think this demonstrates why it's important to turn the copyright of your copylefted code over to the Free Software Foundation.
The Free Software Foundation has a very respected attorney for dealing with violations of the GPL but it can only enforce the GPL if the copyright is held by the Free Software Foundation.
The authors of Everybuddy are very fortunate in this case that the infringers cooperated. What if they hadn't? What would you do? Sue in an Indian court?
Once you have collected the details, you should send a precise report to the copyright holder of the packages that are being misused. The
copyright holder is the one who is legally authorized to take action to enforce the license.
It may do no good to turn the copyright over to the FSF after releasing your code, if an earlier version of your code will serve the infringer's needs - they'll be infringing on your earlier license, not the FSF's later one that they didn't use. That's why you should attend to this right away.
I don't see it in the FSF's pages, but I read that when those folks cracked that censorware program and released it under GPL, that they thought they were off the hook after they settled with Mattel because the code was copyrighted by the Free Software Foundation, so it wasn't their code to withdraw from circulation. But the problem was that they had not sent the FSF a legal written document transferring the copyright. I read in "Legal Care for Your Software" that copyright can only be transferred in writing, I think it may be insufficient to simply state the copyright holder in the source files if it is someone other than yourself.
Slashdot Mirror
But as for state machines, in fact state machines were what I was writing all the times I have written my own scheduler, but I strongly suspect having a kernel developer write a thread scheduler once and have everyone use it is far more efficient than to have every application developer write their own state machines.
While the kernel can deal with it, user-space state machines can't deal with the problem of blocking on a resource. This can only be handled by a kernel scheduler. On a preemptive multitasking but single-threaded system, you achieve efficiency by scheduling a new process when a process blocks; one a multithreaded system, you can schedule a new thread, which may be any thread in any process (or to use the BeOS term, a "team").
There's no way to handle this with user-space threads of any kind, whether they're implemented as some kind of state machine or faked up to look like real preemptive threads.
Michael D. Crawford
GoingWare Inc
But as a result of many horror stories from my own experience and that of my friends, I decided to stop dealing with them and I explain why publicly in Important Note to Recruiters and Contract Agencies.
To make it easier for other consultants to not have to deal with recruiters, I wrote Market Yourself - Tips for High-Tech Consultants.
To directly address the question, though, I think ethics are of the highest importance in the work of a consultant, and are probably the most important guide for you to follow, more important than writing good code. You at least have the hope of debugging bad code.
The question goes both ways though, clients are occassionally unethical and many clients who wouldn't think of screwing you if you were a full-time employee would be happy to short you for weeks of pay earned as a consultant.
You have to protect yourself, start early by finding a good attorney before taking on work - certainly before trouble starts - and have your attorney review all your contracts before you sign them.
Also trust in your feelings and don't do business with someone you feel is not ethical. It's just not worth the heartache.
Michael D. Crawford
GoingWare Inc
I'm pretty perplexed why Linus should think that thread programming is harder. This makes me wonder whether he's ever really tried in a serious way - a lot of programs are much easier to write when you use threads.
Just ask anyone who's had to write their own task scheduler for private use within one program, as I have when writing my Raindrops screensaver for the Mac, or when I wrote a test tool for MacTCP, that enables up to 64 TCP streams and UDP pseudo-streams between any combinations of a Mac on a network - I wrote that in 1990 at Apple, well before threads were available on the Mac OS and it was a pain.
But personally I don't see the point of having threads not implemented in the kernel. It means all your threads block if any of them do. That's the point of kernel threads. I think solaris threads work the way they do because the solaris schedular isn't so efficient, so it's more efficient to multitask some of your threads in user space.
This also relates to why multithreaded programs are more efficient and responsive than single threaded ones, even on uniprocessor machines. If your process blocks to await a resource in a single threaded program, you get to wait while time is wasted. This is OK for servers where avergage efficiency is the main concern but is not appropriate for interactive or near-real-time use.
With a multithreaded program, if one thread blocks for a resource, another thread can run, and if it's another thread in the same program your program's utilization of the CPU will be greater. This is especially the case if the kernel is preemptible and multithreaded, as it is on the BeOS and can be with experimental patches on the Linux kernel.
For an experiment, using the same machine, compare the interactice responsiveness of Windows, the BeOS and Linux, and I can assure you that even for uniprocessor machines the BeOS will be much more efficient, because all native BeOS programs are multithreaded. Threading is available on Linux and Windows but are not yet that widely used.
Michael D. Crawford
GoingWare Inc
Check out Ludwig van on Advogato and also developer Cody Russel's page there.
Michael D. Crawford
GoingWare Inc
It's just a proposal so far, I want to find a few developers to collaborate with before I start, but I'm also writing and posting articles there on the general topic of making Free Software of better quality and also testing the kernel in particular.
While this doesn't provide an answer to your question, some of what I plan would be useful to think about in selecting a bugbase.
The number one feature that I wished to have in bugbases at companies I've worked at in the past was for the user to have an ability to describe a configuration of a machine and then give it a name in the database.
Each user would have some number, one or many, preset machine configurations, particularly the hardware configurations in my case, and the components of these configurations would be drawn from a database describing every piece of hardware that could conceivably be placed in a Linux box.
Then when you go to report a bug, you log in (so your contact info and presets are available), select the config that has the bug, and describe it.
For the Linux kernel, you'd paste in or upload your .config file (this is created by the kernel configuration process).
Then kernel developers could do boolean searches, say "find me all the crashes involving TCP that had a WhizzyNet card installed but in which PPP is not configured in the kernel".
The other part of this is that it is meant to be easy to use for both those who test new kernels (to report bugs) and kernel developers (to research bug reports).
I got the idea to develop the database after subscribing to the linux-kernel mailing list to resolve a bug on my laptop during the 2.4.0-test series kernels.
I was able to work with the list to resolve the bug and see that the patch stuck in the kernel sources, but I felt that many people who might otherwise like to contribute to testing new kernels might be put off by the process of dealing with the list - it's not sufficient simply to report a bug, sometimes patches don't "stick" and you've got to hang around until you're sure your bug stays fixed, but linux-kernel has one of the highest amounts of traffic of any internet mailing list.
Michael D. Crawford
GoingWare Inc
Tell that to the psychiatrists and psychologists who treated me in the psychiatric hospital where I spent the summer of 1985, when I was diagnosed with schizoaffective disorder.
Or to my insurance company at the time, who sprung for the $10,000 treatment.
I know it may be fashionable to claim one is manic depressive, but I can assure you it's unmistakeable in me, and it's not something I would choose. There's no Hell so deep as what I experienced in my college days.
If you don't believe me, maybe you'll at least believe I do some good in my writings. I regularly receive emails in response to my website on Manic Depression from people who've experienced it - either they have it or their loved ones do - and they don't know what to do about it.
There is a lot of information available about the illness but because of the stigma which is perpetuated by ignorant people such as yourself, not a lot of folks have the confidence in themselves to speak out about their experience publicly as I do, occasionally here on Slashdot and permanently on that website.
I speak out in part because of my confidence in my position in my career and also my belief that speaking out on what is right is one of the most important things you can do, even when it comes at great personal cost.
The people who write to me tell me the fact that I speak out personally as I do makes a tremendous difference in their lives. It has happened many, many times that someone has written to me to tell me that I'm the first person they've shared the fact that they think they're mentally ill, or contemplating suicide.
You should know that that page receives about 3,000 hits a month, and I recieve several emails a day from people looking for advice, about half of them people who have the illness and the others from people who are closely involved with a sufferer and don't know how to deal with them.
By the way, I'm happily married. I first met my wife three years ago, and we were married in St. John's Newfoundland on July 22, 2000. We just bought our first house together, in Midcoast Maine, and moved into it last weekend.
Michael D. Crawford
GoingWare Inc
I didn't really know how to program, I knew a little FORTRAN, C and Basic from doing data analysis during summer jobs, and I didn't really like it all that much. I used to really have to struggle to spend several weeks writing a 500 line program, and I'm sure I'd be embarrassed if I had to look at the source code to those programs today.
I figured I'd program for a while because it paid the rent (I was making $20k a year doing Sun administration and writing image processing software), but when I figured out what I really wanted to do for a living I'd quit programming and get a real job.
That was in 1988. Then some consultant visited and installed GNU Emacs on our machines (two Sun 3/160's, one diskless, both with terminals and no workstation monitor, but with frame grabber cards and NTSC color monitors). He explained about the GNU manifesto.
I thought it was pretty cool but didn't see it affecting me personally in a big way. I was mostly annoyed that I had to wait up while the consultant installed the software on what was supposed to be my day off while a ladyfriend was visiting from away.
Then my friend Jeff Keller, who went to MIT for a while and vaguely knew Richard Stallman, spent an evening with me singing the praises of Emacs. What I really wanted was VI with macros you could program to include conditional branches, and he said it had all though and much much more.
So I learned to actually use Emacs, and soon learned that it was quite extensible, but it wasn't made too clear how to extend it. The online manual was useful mainly to people who already knew what they were doing.
So I read the source code. One thing I was interested in doing was writing C functions that were callable from Emacs lisp as lisp functions. There are many such functions built into Emacs (usually for performance) and you can add your own. There's this big DEFUN macro that even makes the C API look like Lisp.
I learned that and a lot more. I learned what an eloquent statement of software architecture Emacs is.
I learned that there really was something worth my while doing in the way of software.
I wanted to write a program like that someday. Not another big editor, but a program that would someday strike other young programmers the way Emacs struck me.
During the course of reading the source code, one day I stayed at my terminal 24 hours straight, arising only to get coffee and use the restroom, not even eating. I only realized how much time had passed when I started to fall asleep.
That was when I started to take programming seriously. I began to put serious effort into studying programming, and studying it deeply.
For example I would read Knuth's The Art of Computer Programming on the bus on the way to work and I would stay up all night after work learning to program better on my Macintosh at home.
For many years I selected all of my jobs based mainly on what I could learn from them.
I've become a very skilled programmer. You can see this from my consulting business website, my resume (on my resume the place where I first encountered Emacs is the Programmer job at Verde Technologies) and my programming tips pages.
So in a very direct and profound way I owe it all to Richard Stallman and Emacs.
I still haven't written my great program yet. I don't even know what it will be. One project I've worked on peripherally is the ZooLib cross-platform application framework and a project I've just started up but not gotten too far with yet is the Linux Quality Database.
I did finally get my B.A. in Physics, from UC Santa Cruz, but only after being out of school working at a programmer for a number of years.
Michael D. Crawford
GoingWare Inc
Yes, this is the list you can submit your name and address to indicate that you don't want to receive unsolicited commercial postal mail. And to some extent it will cut down on certain types of regular junk mail.
But my old boss at Working Software, Dave Johnson, who wrote the chapter on direct mail in The High-Tech Marketing Companion, says that the Mail Preference Service has the very highest response rate of all - for certain kinds of product offers.
(For a long period of time Working Software made most of its sales through direct mail, and Dave became quite an expert on direct mail. This was after he nearly went bankrupt listening to "channel people".)
What kind of product offers sell through this list?
Studded dog collars, burglar alarms, personal security devices, gun magazines and in general products that are aimed at people who are concerned with personal security and just want to be left alone.
Being on the DMA opt-out list doesn't actually prevent you from receiving mail. Instead, members who care to bother (usually because they don't want to waste money sending mail to people who won't respond) get the list periodically and use it to prune their in-house lists. So for lists whose owners bother to go to the trouble, you will be taken off some lists.
But studded dog-collar vendors just take the list and print up mailing labels!
Michael D. Crawford
GoingWare Inc
Maybe the people M$ suspected were up to something didn't have their own resumes on the web, but suppose a low-level employee at their company did, and listed the ex-microsoft execs as references.
I have been told by clients who have actually signed with me to do paying work for them that they hired me because my resume on the web matched the name of some previous employer of mine they were searching for. In those cases, the clients figured that an engineer who worked at these previous companies would be qualified to do their work.
Got your own resume on the web? Does your hosting service give you full logs, not just hit count summaries? Grep out your resume URL and look at the referring pages - in many cases the referring page URL will include the search keywords someone used to find your resume. It's pretty interesting to see what people are entering and then clicking on my resume as a result.
Michael D. Crawford
GoingWare Inc
Don't have a fax? Maybe the local copy shop downtown will allow you to receive faxes for a fee. I've known drug stores where I received faxes regularly.
Read the contracts thoroughly. You want my best advice, no matter what it says, take it to an attorney.
Make it clear to your new employer that you're going to be paying to have an attorney review their contracts, and if they make any changes later you're going to have to consult an attorney again. If they give you a hard time about this let them know that you'll be billing them for the additional expense.
Now, it's very important that you've asked up front for your contract. If the contract hasn't arrived before the day you're to start work, call in and say you can't show up because you haven't had an opportunity to see the contract. Tell them, as earnestly as you can, that a contract is a serious commitment between you and them and you wouldn't dream of signing one without reading it first, at home, without them watching you, and having access to an attorney.
If they give you a contract you haven't seen before when you arrive, inform them that you'd made clear your desire to review any contracts ahead of time and walk out.
If they're really worth working for, they'll get you the contract to review. If you're the kind of person for which it would really be an issue to have a non-compete agreement at all (exec, high-level engineer etc.), then they're not just going to let you go work for someone else after you've made it clear they insulted you and tried to deprive you of your right to make informed decisions about your own future.
Did you get the job through a headhunter or contract employment agency? You should know that they're paying, at minimum, 30% of your first year's salary for the headhunter's fee to hire you, or 30% of your hourly rate to the agency for the entire time you work (not just the first year). Consider how valuable you must be if they're paying this much for you, so really it's not that much to ask that you be completely informed before showing up to work.
To make it crystal-clear: Your decision to accept the job or contract should be based in part on full information about the contracts you will be signing. Do not make this decision until you've read the contract. Tell them that acceptance of their offer letter is conditional on reading their contract first.
I'm not kidding about this and this is not a troll. I really mean this and I'm speaking from hard experience. I've been a software engineer for thirteen years now and have been running my own independent consulting business for almost three years (by independent, I mean I don't work through agencies, so I'm on my own when I sign a contract).
You think you really need a job? Tired of flipping burgers - or think that's what you'll be doing if you don't sign? Thimk of the months or years of misery that will result if you make a rash decision right from the start. Not only will you regret it, but you may lose tens of thousands of dollars as a direct result of taking this job, rather than turning it down and finding another, or negotiating a better contract.
There's something called "opportunity cost" that many managers know about but not a lot of regular employees. It is the cost of losing the opportunity to do something else. You may think you've gained by getting a job, but you may have lost far more by not turning it down and taking a different one, or getting rid of that noncompete so that when this job is done (and no high-tech job lasts forever) you can get the rewarding job of your choice.
Michael D. Crawford
GoingWare Inc
The developers were out their shipping many fine, supported commercial products that worked very well - Gobe, Steinberg was onboard developing audio applications, Thomas Dolby spoke at the developer conference about supporting Be with his audio software (he was wildly enthusiastic).
No, the incompetence was not the developers. It was Be's press relations and marketing (or to be more precise, the lack thereof). Be's technical staff is highly competent, and they produced a fine product, something Linux developers would do well to study because of it's ease of installation, smooth and trouble-free performance, and clean integration. It's also a joy to program.
My wonderfully positive experiences with using the BeOS and my sadness at having to leave the platform behind in order to have currently supported hardware and software is another component of what motivates my effort to improve the quality of Linux and Free Software in general.
But Be's business management was always, and still is, incompetent. There was never any effort made to convince people to make the move to the BeOS - they were always waiting for more applications to appear, even when there were plenty of applications for the average user to do almost all their daily work.
Even though almost anyone who ever laid eyes on the system became an instant convert - and many of those became evangelists, such as I especially after I got it running on my laptop, Gassee never felt it was ready for the mass market and kept holding it back from the people who could have done him the most good - the public who was eager to try something better if only they knew it existed and where to get it.
For example, I never once saw a banner ad on a website that was run on a non-Be related site directing the user to Be's website or BeDepot (Be's now-defunct eCommerce site).
And speaking of BeDepot - Gassee very early on spoke repeatedly about the problem of penetrating software channels for new developers, especially developers for new platforms like the BeOS, and how Be was going to be very advanced in its thinking by offerring software for sale via download at BeDepot.
And we all have seen what has become of eCommerce in general - it's the only way you can get many products for mainstream platforms - but BeDepot was one of the most ineptly run operations I have ever seen.
BeDepot's ineptitude was murder to developers for whom it was the only sales channel - so you've got a newly released product? What do you do when it takes three months to appear on the website because they're either too understaffed to deal with it or too incompetent to just upload the damn file? I've heard lots of complaints from developers who had to wait months to receive their contract from Be that would enable them to sell on BeDepot.
It's not like it's so hard to run a good online Be software site, as demonstrated by the folks who founded BeBits, in part to make up for BeDepot's and BeWare's inadequacies. (BeWare was for free downloads).
If you won't listen to me because of Be's lack of character, listen to me because of Be's demonstrated incompetence, and know that if they couldn't keep an eCommerce site with a few dozen titles running efficiently, how are they going to provide adequate support for a bunch of disparate embedded hardware vendors?
By the way, you should note that many of Be's most prominent engineers left when Be made the decision to drop the desktop and move to Internet Appliances, folks like Dominic who wrote the BFS Journaled Filesystem (with indexed file attributes - there's a read-only version of BFS you can get for Linux, that I'd like to make read-write, and would make a great addition to the system).
Michael D. Crawford
GoingWare Inc
After years of silver-tongued promises of support from Be for such things as marketing support for developers, Be never effectively marketed it's own BeOS desktop operating system and then announced that it was ceasing development of it, except to the extent necessary to support it's development of the BeIA internet appliance platform.
This means, for example, that the current rev of the BeOS won't install on my PC because I use an Ultra160 SCSI controller - Be doesn't see these as necessary for support of it's development platform.
And you can forget about marketing support for those of use who listened to Be's promises and actually shipped commercial products for the BeOS.
While I may have lost a lot of time and effort coding spellcheckers that could have been better spent elsewhere, Be developers BeatWare and Adamation lost millions, and are only able to survive because they ported to Windows or Mac.
My essential point is that Be is not to be trusted by Sony when times get difficult for it. There was a time when folks like me, Adamation and Beatware were all Be had. If Be can screw us, why can't they screw Sony?
Michael D. Crawford
GoingWare Inc
This lack of integrity goes to the heart of Be's corporate culture, and is in particular exemplified by the attitudes and practices of Be's CEO, Jean Louis Gassee.
Read about the experience of one of Be's formerly most-loyal and most enthusiastic business partners in:
-
The Valley is a Harsh Mistress
Note: the page gives the URL's for a number of alternative vendors for Internet Appliance operating systems, many of which are open source (an important consideration in a market where your vendor may abandon you). Also see the Embedded Linux Journal.I have no doubt that BeIA has tremendous technical advantages. See what this once-ardent BeOS developer is doing to harden the competition for Be Inc. at The Linux Quality Database and Freeing the Developer from OS Vendor Shackles.
Michael D. Crawford
GoingWare Inc
- http://linuxquality.sunsite.dk
So far it's just a proposal. In the short term there will be resources on testing strategies, as well as tips on writing quality code. I the long run will come a nice web form for reporting bugs in a way that will be especially meaningful for kernel developers (capturing and searching hardware configuration and kernel config options).Michael D. Crawford
GoingWare Inc
-
The Web is the Ultimate Copy Protection
Even with a proprietary closed-source program, you at least still have the distribution media so you can still reinstall your application if the original publisher stops selling the program.But with a web-based application, the publisher keeps the executable binary, and even with copylefted code, the GPL's source code distribution requirement does not come into play because the binary is never distributed.
This will ultimately lead to disaster for ordinary users as web applications make the move from news, shopping and entertainment to functions that people depend on in their daily lives or businesses.
Michael D. Crawford
GoingWare Inc
My manager let it slip then asked me to keep quiet about it.
Been through a couple layoffs at Apple. Read about what I think about Apple's management. But don't think I'm still a fan of Be, Inc. - read about what I think we ought to do to all operating systems vendors.
Michael D. Crawford
GoingWare Inc
And it always seems to happen when I'm composing something in a web form like here on Slashdot, so I end up losing it, so I immediately hit the submit button when it started happening while composing the above.
Working towards improving this situation is The Linux Quality Database.
Michael D. Crawford
GoingWare Inc
If is very easy to figure out your nearest mirror:
Determine the two-letter country code for the nation you reside in. This will be the same as the top-level national domain, ca for Canada, uk for United Kingdom, ch for Switzerland and so on.
If you're using ftp, the server name starts with ftp. If you're using http (a web browser), then it starts with www.
Put the country code in the middle. Add .kernel.org at the end.
Michael D. Crawford
GoingWare Inc
Distros update their versions for lots of different reasons, often because a number of new user programs are available, a new XFree86 version is available that is needed for new video cards, or sometimes for purely marketing reasons, as when Slackware jumped from 4.0 to 7.0 in order to reach marketing parity with Red Hat - Slackware generally used a conservative numbering scheme, but Red Hat advanced the numbers rapidly, even though all the distros are based on the same stuff, but lots of people gave the Slackware folks the message they didn't want software they perceived as out of date.
So no, it's fine to use the 2.4.0 kernel with Mandrake 7.1. The version of the kernel bears no explicit relation to the version of the distribution.
By the way, if you wonder where most of the rest of the programs on your distro come from, you have the fine folks at The Free Software Foundation to thank. Richard Stallman takes pains to point out the system is more properly called "GNU/Linux", because what Linus developed was a kernel that the already-existing GNU programs could run on.
Michael D. Crawford
GoingWare Inc
I just saw a very confused user posting to linux-kernel wondering where to send a report (he'd figured it out, but wasn't sure).
The correct, new address for bug report submissions is linux-kernel@vger.kernel.org
Same hostname, different domain.
If you're going to work with the new kernel, I highly recommand you browse the mailing list. But linux-kernel has one of the highest volumes of any internet mailing list so you probably don't want to actually subscribe (you don't have to subscribe to post, unlike many mailing lists).
Instead, read the list off an archive server. There are many of them. This search at Google will find you an archive
Michael D. Crawford
GoingWare Inc
I got in before the Slashdot effect myself - I'm running 2.4.0-prerelease-ac5.
They have this time at Apple in the development of new hardware called freezing the ROMs. You have to do it so you can get them into production to include in the new machines. But development of ROM software doesn't stop, it just takes the form of RAM-based patches...
Michael D. Crawford
GoingWare Inc
So just yesterday I wrote:
-
Why We Should All Test the New Linux Kernel
on Advogato.It's got some helpful, practical tips on downloading and building a kernel from sources as well as providing feedback to the kernel developers. All of the information in the article is available somewhere, but when I first began testing with 2.4.0-test1, I found some things difficult to figure out, so I felt that it would be nice to put what I learned all down in one place.
This is part of an overall effort to improve the quality of Free Software. Another part of the effort is the just-founded Linux Quality Database, so far just a proposal - contact me at crawford@goingware.com if you want to help.
And finally, some interesting trivia for you:
After a long beta testing period and many delays, Windows 2000 shipped with 64000 documented bugs of which 25000 were considered serious by Microsoft itself (the figures are quoted from memory, I might be a bit off). The Windows trade press reported that the opinion was widely held by IT managers that one should not install Windows 2000 on any machine until a few service packs had been released, with Windows 2000 server not being considered ready for use until much later than the desktop user version.
The BSD/Mach based Mac OS X, derived from the NeXT operating system and now in beta testing, is Apple's first operating system to support protected memory that is expected to be widely used. (Another was A/UX, Apple's Unix port, but it served only a niche market).
Apple has been trying for ten years, longer than Linux has been in existence, to write a modern operating system. The first was Pink, renamed Taligent when Apple collaborated with IBM on it. In the end all that came of it was a little-used object-oriented programming framework.
The next was Copland, and I don't know why it was never completed exactly, but I was offerred a job as a performance engineer on the Copland project when I worked at Apple, but I turned it down - I didn't tell them this but I had the sense that Copland was a project that would never ship.
And Linus wrote Linux when he was a college student, which combined with the GNU utilities forms the operating system we know and love today.
These huge, well-funded corporations can't get it together to program their way out of a wet paper bag but a bunch of freaks on the Net have written the fastest growing operating system in use today.
Michael D. Crawford
GoingWare Inc
-
http://www.bustpatents.com
You think Aharonian's a hero? You don't know the half of it - he's being sued by an intellectual property holding company for patent infringement. The basis of their suit include such claims as the accusation that Aharonian Read about it in Patent Fight Pending.Michael D. Crawford
GoingWare Inc
In many cases you could change the text to say "Copyright (c) 2000 Free Software Foundation. This program is free software; yadda yadda".
If you're program is a GUI program, it may not be too hard to obscure the generation of this text and show it in a message box.
One thing I suggest though is that if you do this, you split up the code that does it and distribute it all over your program, and include several independent ways it can be done with independent implementations.
Michael D. Crawford
GoingWare Inc
The Free Software Foundation has a very respected attorney for dealing with violations of the GPL but it can only enforce the GPL if the copyright is held by the Free Software Foundation.
The authors of Everybuddy are very fortunate in this case that the infringers cooperated. What if they hadn't? What would you do? Sue in an Indian court?
Please see what to do if you think you see a violation of the GPL, LGPL, or GFDL and note what they say:
It may do no good to turn the copyright over to the FSF after releasing your code, if an earlier version of your code will serve the infringer's needs - they'll be infringing on your earlier license, not the FSF's later one that they didn't use. That's why you should attend to this right away.To make the Free Software Foundation the copyright holder of your package, please see the FSF's info on legal matters and how to apply the GPL license terms to your new programs.
I don't see it in the FSF's pages, but I read that when those folks cracked that censorware program and released it under GPL, that they thought they were off the hook after they settled with Mattel because the code was copyrighted by the Free Software Foundation, so it wasn't their code to withdraw from circulation. But the problem was that they had not sent the FSF a legal written document transferring the copyright. I read in "Legal Care for Your Software" that copyright can only be transferred in writing, I think it may be insufficient to simply state the copyright holder in the source files if it is someone other than yourself.
Michael D. Crawford
GoingWare Inc