Seriously, what's going on with the articles here? "My code is like a Taco"? Is that flying because of CmdrTaco's username?
Nothing new here:
1) Code reuse. Woopdeedoo. The whole industry has invested heavily in many paradigms for reusing code: The reusable library, module reuse, object reuse etc.
2) Stringing Unix commands together is news? Did I just take a Deloriane back to 1955? (Well that's a slight exaggeration. Unix has only been around since the 70s)
Finally, who wants to compare their code reuse to a crappy junk food chain? I'd rather think of myself as a professional that earns commensurate pay than a junk food server who needs to be trained to ask "would you like fries with that?".
Maybe you didn't read the article, but you've totally missed the point. The point is that these duct-taped solutions with shell scripts are often good enough for the first draft of an application. The contrast is with many people trying NoSQL and MapReduce systems because they're the flavor-of-the-day. The point is that the Unix paradigm is powerful enough for the first draft of many systems, and depending on your needs might be enough for the final draft.
Also, you can be anyone you want in the analogy. Maybe you invented Taco Bell. Maybe you're the person who invented the 4-ingredient Chalupa, which probably makes millions of dollars per year. I never thought that I would be a junk food server. That's clearly a job for a script.:)
I understand that many people will want to jump on this man and lambast him for his brash, profanity-laced rhetoric towards members of the media. Please be aware that he will soon lose his job after Yahoo's market share is swallowed up by companies that are more effective at building a user base, maintaining market share, and turning a profit.
It's not easy to face the looming demise of your company and your job with candor, intelligence and restraint. So please be nice to this guy as his company goes down the tubes.
While I agree that folding@home is more useful than seti@home, I think that Rosetta@home. It's also focused on protein folding, but the difference is that Rosetta has consistently outperformed folding@home at the CASP (Critical Assessment of Structure Prediction) competitions. Check out the CASP website to see the raw results. Or, check out a summary from the Baker Lab website. Also, Dr. David Baker (head of the lab where Rosetta has been developed) is very involved in the community of users that run Rosetta@home, check the messageboards on the Rosetta@home site.
Disclaimer: I'm a student in David's lab. But that doesn't mean that I'm wrong, or mindlessly plugging my own Kool-Aid.:) I really believe that Dr. Baker and his lab have a strong chance to solve the protein folding prediction problem.
Whatever project you choose to donate your cycles to in the end, protein science is a cool field with far-reaching implications for humans in general, and the scientists in the field really appreciate your cycles. Thanks to all those who are donating and will donate in the future.
Don't send mixed messages. If you dress up in an outfit where half your ass is out of your skirt and your boobs are almost poping out of your tank top then expect for some guy to treat you like a whore. Now guys, that doesn't mean they are whores. But damn it's confusing.
If I wore a police uniform and some guy came up asking me for help could I say "Hey! Just because I'm dressed as a policeman doesn't mean I am one." No man, that's just confusing. So next time you decided to dress up like a whore, just remember that you may not be a whore, but you're wearing a whore's uniform.
DJ Rat is recording his new mixed tape, which should be released early next Spring. The FCC; however, is not so thrilled, because the mixed tape "is one of the dirtiest fowlest things I have ever heard" said an FCC spokesman. "But it wasn't a suprise I guess, you know- he is a rat after all".
The mix tape would be more fowl if it were written by a duck.
i guess you don't know what stratification means.
I understand the meaning of the word. It generally refers to the separation (conceptual or physical) of a group of objects into their separate components (or strata). It's often used in geology, meteorology and statistics has a context-dependent meaning associated with the environment in which it was used. I'm assuming that the strata you're referring to are disparate software layers, many of which can be GPL. That's not the issue.
The issue is that you didn't address a single point that I made in my post, so I'll break one down for you: showing your boss that use of GPL software increases productivity does not clearly support the argument that your software should be GPL. If I were your boss and you made this argument to me, I would say: "Great! Use all the GPL software you want!" Nowhere in that discussion would I think about letting you release your software as GPL or open-source.
Businesses are driven by bottom-line performance. You can fart around with concepts of software "wanting to" be free all you want, but until you can provide concrete evidence that releasing open-source software benefits your company's bottom line, I wouldn't expect to work on open source software at work.
Other threads have provided some good arguments for releasing OSS. Here's some of the ones I like:
Releasing open-source software can stimulate adoption of a platform, which can later be leveraged by other software products at the company.
Open-source software allows external programmers to add features and fix bugs in software.
Testing can be done at a reduced cost by setting up and monitoring a mailing list or forum.
I'm not saying that you're wrong, I'm just saying that you're not making a compelling case from a business perspective.
What the hell are you talking about? The question isn't "Hey, can you motivate me to write some useful open-source code?" He already has code that he'd like to GPL. How can he convince his boss that this is a good idea?
That is to say, position yourself with GPL/OSS code in such a way that you prove that the open stratification of code, abstract on whatever levels you choose, is worth the effort. Write an app using GPL'ed API's that nobody else could've written, quite so quickly, if a deep and sudden understanding through direct study of contributed frameworks wasn't done first.
While those are some nice multi-syllabic words, you've missed the point. Convincing your boss to let you use use GPL/OSS code is not the same as convincing your boss that you should release your code under a GPL license. It sounds from the OP that he app he's writing is an internally used piece of software, which means that the source doesn't even have to be publicly available (until the app is distributed somehow).
GPL and F/OSS are about code and code is all about what runs, not what someone 'thinks' about it.
I disagree with this as well. Don't you think that the perception of OSS affects how many people will run it and how it will be used? Think about it this way: if I wrote a paper that inspires my boss to allow me to open-source some of my code, doesn't that benefit the open-source community? If I can code open-source software at work, doesn't that result in more open-source software being written? Isn't that a good thing?
I think that code is very important, but I think that a myopic view of source as the end-all be-all of life is disturbingly myopic. Especially for those of us who live life in the real world, with jobs that pay in real money.
Why do you claim that the market research "falsely indicates that the average consumer loves these damn things?" I'd like to hear what evidence you have to the contrary.
I don't quite know what you mean when you claim that Bioperl, Biopython, EMBOSS and BioConductor are more oriented towards servers than stand-alone applications. First of all, servers and stand-alone applications don't divide up the application world into mutually exclusive parts. Applications can be stand-alone and run on a server for example. I've built applications using Bioperl that have a GU interface (take that grammar nazis!), and people are extremely happy with them. So, if you have a Perl guy nearby, I highly recommend talking to them about your problems.
Secondly, translations? Database searches? Sounds like you're doing some very basic Bioinformatics work. Not to say that your research isn't meaningful, just that the problems you're approaching are easily solved by a computational biologist. For example, here's a snippet of Bioperl code that will read in a set of GenBank sequences, translate them and print the results to a new file: my $seqin = Bio::SeqIO->new( -file => 'myseq.gbk', -format => 'genbank' ); my $seqout = Bio::SeqIO->new( -file => '>translated.gbk', -format => 'genbank' ); while ( my $seq = $seqin->next_seq ) {
$translated_seq = $seq->translate;
$seqout->write_seq( $translated_seq ); }
Seems pretty simple, right? There are similar, simple wrappers around BLAST, FASTA and some other common algorithms in computational biology. Check out the Beginners HOWTO on the Bioperl website, it explains Bioperl without requiring previous CS experience. I think it's a good intro, but I also wrote it so I'm slightly biased.
If programming is not your style, check out JEMBOSS. It's a Java-based GUI wrapper for EMBOSS.
The OP said he'd buy Frontpage if it were available for his platform, but he won't spend the meager amount of money to find a decent webhost? That's ridiculous, I think that a (legal) version fo Office would cost at least $100, which should be more than a year of hosting from a decent webhost.
This is amazing for several reasons. First, I think I'd get fired for letting software (that may or may not be working correctly) do a job that is so important and not have any humans checking the work.
You very obviously don't work in the field of proteomics. There's really not a good way for a human to check the algorithm's work without doing more laboratory experiments on the spots. So, having a human checking the work is more detailed (and more expensive) than having a guru glance over the output. The problems that the OP was talking about aren't the kind that a human eye can interpret, and I have some doubts about whether a 2D-gel even has the necessary information to answer some of the questions that he asks. There may be some software vendors that claim that their software can tell you how many different proteins are in a spot, but I wouldn't buy that claim without seeing some very good experimental evidence.
I imagine that scanning protien smears is not much more difficult than finding micrometer sized fractures in the skin of an airplane, or finding hard to see stars using amature telescopes and computer driven camera technology.
I disagree with your imagination. It's not just a problem with identification of spots on the protein gel, the big problem is in interpreting what the spots actually mean. To use your example, it might be like finding micrometer sized fractures on the skin of an airplane skin, and then attempting to pick what sort of object initially caused the fracture (assuming that the fractures are caused by collisions with debris, rather than some defect of the manufacturing process).
You are correct in that rather than AI, the OP should be looking for image analysis algorithms. After taking a course in computer vision last fall, I found that even object recognition algorithms aren't as refined as I thought they would be.
A spokesman for the Recording Industry Association of America said his organization had not reviewed the software, but said that in principle it was disturbed by the idea. "We remain concerned about any devices or software that permit listeners to transform a broadcast into a music library," RIAA spokesman Jonathan Lamy said.
The RIAA and XM are both busy figuring out if any copyright laws and user agreements have been broken.
Nowhere in the article is there any mention of fair use rights or the legality of this sort of software. The RIAA is obviously very concerned about this, as it would definitely affect their willingness to release entire albums over the air. Blah.
This just in: MIT has released a windowing system that allows for multiple users on a time-sharing system to use GUI applications across a network.
Seriously though, this is cool. It leverages the often unused computing power of workstations that are traditionally used as thin clients so that processing can be done. Definitely a cool project using cheap hardware and free software.
Imagine a world where PHBs can turn their office wall into a window onto any cube. Zero privacy.
Yeah, it would be very similar to a world full of ads for X-10 cameras and high availibility of office security cameras. I wouldn't want to live in a world like that. I mean... riiiiiiiiiiiiiiiiight.
Why do you think that every person goes right? For the exact same reason that people play games in the first place: excitement. Many people who play games are trained to go towards the machine guns, lasers, mirrors and battles in video games because that's traditionally wbere the action is in games. Gamers have been trained to do things like that because that's what games have taught them to do.
I like some of the ideas that are put forth in the article, but I think that people will gladly come up with new and interesting ways to succeed in games as the physics and AI models become increasingly complex.
These are all very good solutions. More ideas include high availability network stuff such as UCARP or HA-Linux. Many of these depend on specific details of your setup, these recommendations are definitely biased towards Unix/Linux systems. In my experience, having backup systems for your important is a crucial idea, having employee downtime as a result of system failure is a nightmare for any company, especially a small business.
A mysteriously soldered VRM sounds a bit odd, you might think about going with a rock-solid hardware platform if/when you upgrade your current infrastructure. I've had great success with a Sun E450 for the last two years, it's a rock-solid piece of equipment that's required absolutely no hardware maintenance.
It's difficult to give you more concrete recommendations without knowing more about your setup. Things like hot database backups can be tricky, especially if your IT staff isn't experienced in database backup solutions. You might find value in discussing some ideas with a consulting company. The parent post seems to know a lot about this, and there are certainly other companies out there.:P
Slashdot Mirror
Seriously, what's going on with the articles here? "My code is like a Taco"? Is that flying because of CmdrTaco's username?
Nothing new here: 1) Code reuse. Woopdeedoo. The whole industry has invested heavily in many paradigms for reusing code: The reusable library, module reuse, object reuse etc. 2) Stringing Unix commands together is news? Did I just take a Deloriane back to 1955? (Well that's a slight exaggeration. Unix has only been around since the 70s)
Finally, who wants to compare their code reuse to a crappy junk food chain? I'd rather think of myself as a professional that earns commensurate pay than a junk food server who needs to be trained to ask "would you like fries with that?".
Maybe you didn't read the article, but you've totally missed the point. The point is that these duct-taped solutions with shell scripts are often good enough for the first draft of an application. The contrast is with many people trying NoSQL and MapReduce systems because they're the flavor-of-the-day. The point is that the Unix paradigm is powerful enough for the first draft of many systems, and depending on your needs might be enough for the final draft. Also, you can be anyone you want in the analogy. Maybe you invented Taco Bell. Maybe you're the person who invented the 4-ingredient Chalupa, which probably makes millions of dollars per year. I never thought that I would be a junk food server. That's clearly a job for a script. :)
It's not easy to face the looming demise of your company and your job with candor, intelligence and restraint. So please be nice to this guy as his company goes down the tubes.
Disclaimer: I'm a student in David's lab. But that doesn't mean that I'm wrong, or mindlessly plugging my own Kool-Aid. :) I really believe that Dr. Baker and his lab have a strong chance to solve the protein folding prediction problem.
Whatever project you choose to donate your cycles to in the end, protein science is a cool field with far-reaching implications for humans in general, and the scientists in the field really appreciate your cycles. Thanks to all those who are donating and will donate in the future.
If I wore a police uniform and some guy came up asking me for help could I say "Hey! Just because I'm dressed as a policeman doesn't mean I am one." No man, that's just confusing. So next time you decided to dress up like a whore, just remember that you may not be a whore, but you're wearing a whore's uniform.
(Paraphrased from a David Chappelle standup bit.)
DJ Rat is recording his new mixed tape, which should be released early next Spring. The FCC; however, is not so thrilled, because the mixed tape "is one of the dirtiest fowlest things I have ever heard" said an FCC spokesman. "But it wasn't a suprise I guess, you know- he is a rat after all".
The mix tape would be more fowl if it were written by a duck.
The issue is that you didn't address a single point that I made in my post, so I'll break one down for you: showing your boss that use of GPL software increases productivity does not clearly support the argument that your software should be GPL. If I were your boss and you made this argument to me, I would say: "Great! Use all the GPL software you want!" Nowhere in that discussion would I think about letting you release your software as GPL or open-source.
Businesses are driven by bottom-line performance. You can fart around with concepts of software "wanting to" be free all you want, but until you can provide concrete evidence that releasing open-source software benefits your company's bottom line, I wouldn't expect to work on open source software at work.
Other threads have provided some good arguments for releasing OSS. Here's some of the ones I like:
- Releasing open-source software can stimulate adoption of a platform, which can later be leveraged by other software products at the company.
- Open-source software allows external programmers to add features and fix bugs in software.
- Testing can be done at a reduced cost by setting up and monitoring a mailing list or forum.
I'm not saying that you're wrong, I'm just saying that you're not making a compelling case from a business perspective.That is to say, position yourself with GPL/OSS code in such a way that you prove that the open stratification of code, abstract on whatever levels you choose, is worth the effort. Write an app using GPL'ed API's that nobody else could've written, quite so quickly, if a deep and sudden understanding through direct study of contributed frameworks wasn't done first.
While those are some nice multi-syllabic words, you've missed the point. Convincing your boss to let you use use GPL/OSS code is not the same as convincing your boss that you should release your code under a GPL license. It sounds from the OP that he app he's writing is an internally used piece of software, which means that the source doesn't even have to be publicly available (until the app is distributed somehow).
GPL and F/OSS are about code and code is all about what runs, not what someone 'thinks' about it.
I disagree with this as well. Don't you think that the perception of OSS affects how many people will run it and how it will be used? Think about it this way: if I wrote a paper that inspires my boss to allow me to open-source some of my code, doesn't that benefit the open-source community? If I can code open-source software at work, doesn't that result in more open-source software being written? Isn't that a good thing?
I think that code is very important, but I think that a myopic view of source as the end-all be-all of life is disturbingly myopic. Especially for those of us who live life in the real world, with jobs that pay in real money.
Why do you claim that the market research "falsely indicates that the average consumer loves these damn things?" I'd like to hear what evidence you have to the contrary.
Secondly, translations? Database searches? Sounds like you're doing some very basic Bioinformatics work. Not to say that your research isn't meaningful, just that the problems you're approaching are easily solved by a computational biologist. For example, here's a snippet of Bioperl code that will read in a set of GenBank sequences, translate them and print the results to a new file:
my $seqin = Bio::SeqIO->new( -file => 'myseq.gbk', -format => 'genbank' );
my $seqout = Bio::SeqIO->new( -file => '>translated.gbk', -format => 'genbank' );
while ( my $seq = $seqin->next_seq ) {
$translated_seq = $seq->translate;
$seqout->write_seq( $translated_seq );
}
Seems pretty simple, right? There are similar, simple wrappers around BLAST, FASTA and some other common algorithms in computational biology. Check out the Beginners HOWTO on the Bioperl website, it explains Bioperl without requiring previous CS experience. I think it's a good intro, but I also wrote it so I'm slightly biased.
If programming is not your style, check out JEMBOSS. It's a Java-based GUI wrapper for EMBOSS.
Cheers and good luck.
The OP said he'd buy Frontpage if it were available for his platform, but he won't spend the meager amount of money to find a decent webhost? That's ridiculous, I think that a (legal) version fo Office would cost at least $100, which should be more than a year of hosting from a decent webhost.
Or did you mean to say TB?
You very obviously don't work in the field of proteomics. There's really not a good way for a human to check the algorithm's work without doing more laboratory experiments on the spots. So, having a human checking the work is more detailed (and more expensive) than having a guru glance over the output. The problems that the OP was talking about aren't the kind that a human eye can interpret, and I have some doubts about whether a 2D-gel even has the necessary information to answer some of the questions that he asks. There may be some software vendors that claim that their software can tell you how many different proteins are in a spot, but I wouldn't buy that claim without seeing some very good experimental evidence.
I imagine that scanning protien smears is not much more difficult than finding micrometer sized fractures in the skin of an airplane, or finding hard to see stars using amature telescopes and computer driven camera technology. I disagree with your imagination. It's not just a problem with identification of spots on the protein gel, the big problem is in interpreting what the spots actually mean. To use your example, it might be like finding micrometer sized fractures on the skin of an airplane skin, and then attempting to pick what sort of object initially caused the fracture (assuming that the fractures are caused by collisions with debris, rather than some defect of the manufacturing process).
You are correct in that rather than AI, the OP should be looking for image analysis algorithms. After taking a course in computer vision last fall, I found that even object recognition algorithms aren't as refined as I thought they would be.
Not to sound pompous, but is this really that high of a standard? I'm not trolling, I just really don't know.
A spokesman for the Recording Industry Association of America said his organization had not reviewed the software, but said that in principle it was disturbed by the idea. "We remain concerned about any devices or software that permit listeners to transform a broadcast into a music library," RIAA spokesman Jonathan Lamy said.
The RIAA and XM are both busy figuring out if any copyright laws and user agreements have been broken.
Nowhere in the article is there any mention of fair use rights or the legality of this sort of software. The RIAA is obviously very concerned about this, as it would definitely affect their willingness to release entire albums over the air. Blah.
I'll bet that most Slashdotters would be denied service from the average girl. :P
*dodges rotten tomatoes*
:P
Your move Bill.
Yes.
Seriously though, this is cool. It leverages the often unused computing power of workstations that are traditionally used as thin clients so that processing can be done. Definitely a cool project using cheap hardware and free software.
Note to the editors: What do I pay you people for anyway?
:P
Yeah, it would be very similar to a world full of ads for X-10 cameras and high availibility of office security cameras. I wouldn't want to live in a world like that. I mean ... riiiiiiiiiiiiiiiiight.
:P
I like some of the ideas that are put forth in the article, but I think that people will gladly come up with new and interesting ways to succeed in games as the physics and AI models become increasingly complex.
You could even blend your two solutions, so that you could use a rule like "Block all windows spawned by forbes.com that use this javascript code."
A mysteriously soldered VRM sounds a bit odd, you might think about going with a rock-solid hardware platform if/when you upgrade your current infrastructure. I've had great success with a Sun E450 for the last two years, it's a rock-solid piece of equipment that's required absolutely no hardware maintenance.
It's difficult to give you more concrete recommendations without knowing more about your setup. Things like hot database backups can be tricky, especially if your IT staff isn't experienced in database backup solutions. You might find value in discussing some ideas with a consulting company. The parent post seems to know a lot about this, and there are certainly other companies out there. :P
Best of luck to you.