Are these deals with large manufacturers or "small" companies? On what forums do these deals take place? I'm not a gearhead, just interested in the economics of it all.
Yeah, when I read the headline "divide an image into discrete patterns", I was reminded of the 'Applied Probability Course' offered by ArsDigita.org. One of the lecturers appearantly did a PhD thesis on this topic. So, I was kind of impressed that some father would be interested in going to such lengths for his kid. (IIRC, there were only 5 lectures for the course, so you could certainly watch them if your interested in this subject.)
Tip: Most 'hard' problems have really nifty approximate solutions that are good enough for most applications (i.e. problems not in AI/automation/verification).
P.S. I think the algorithm using a 'histogram of top N colors' is pretty cool for a automated paint-by-number pattern generator. Seems like that could be a cool perl hack. Sure, you might have to throw a way some of the more 'complicated' images that have too many shades of a color (or other similar low probability scenarios). Maybe you could examine the difference in RGB values of colors in the pallete and examine the frequency in which these colors are used.
if it's any consolation, most logicians are so nitpicky, they actually annoy other types of mathematicians. It kind of makes you wonder what conversations with Godel was like?
Most restaurant owners aren't IT savvy. Especially if it involves exotic hardware, a non-windows OS, and sparsely documented software. Heck, most restaurant managers have problems using the windows machine in the back. From what I've seen (which admittedly isn't much) most restaurants hire a technician to maintain their POS setup. I hope the owner is paying the article poster. Frankly, he's going to need support no matter what.
If you want to write viruses and easy-to-use hacking tools -- yes. Look, Perl is the glue for *nix. However, thanks to Microsoft's object models, VB/VBA/Visual Studio Tools for Office is the probably more the Window hackers tool of choice that many would care to admit. Of course, how many respectable 'hackers' hack Windows?
A contract is only as good as your ability to enforce it in court. While legally you are correct in saying that a computer transaction can be made into a legally binding agreement, producing evidence that joe sixpack @ 111.111.111.111 is really who he says he is and collecting on that contract may be more difficult. Do I really want to bring a computer forensics expert or my sysadmin to court? Plus, tying a face, identification, and a 'physical presence' to a legal contract only adds to the legitimacy.
If you're bringing in $100,000+ per month in sales, you probably should get your own site. EBay is great for the 'small-time' or 'early-stage' seller. It's a robust engine that can handle heavy traffic and activity and is reasonably secure. However, if you've got sales as high as mentioned in the article, I can't imagine how EBay could provide an adequate solution for you. If you sell big ticket items, you probably want some binding contractual arrangement. On the other hand, if you have an incredibly high volume of sales, you probably want to better cater the shopper's experience and would have an E-commerce site of your own. Not to mention the fact that you'd need some internal system (even an Access database) to track all of your sales and shipping data. Maybe EBay's API allows you to do some of these things. I'd certainly be happy to hear about anyone experiences with it.
If it's any consolation, the matrix violates the laws of thermodynamics. Plus, the human body is a grossly inefficent battery. Mr. Sun is what helps the little 'open system' that is Earth go. In other words, the robots would probably kill us and use us a fertilizer instead of an energy source. So don't worry!
Let's face it. Public schools are large lumbering institutions. I learned very little there. Sure, the governor's school and advanced courses are usually pretty good, but the average public school education is filler at best. The modern day educator is afraid of legal suits, sexual harrassment charges, excessively disrespectful behavior by students (you know it's true), and a largely hostile collection of parents. Schools respond by imposing draconian rules and excessive punishment for relatively petty offenses. It's truly an administrative mess. Why congress isn't debating THIS 24/7 is truly mystifying. Unless I'm dead broke, I don't plan on sending my kid to a public school.
If anything, I want these legislators regulating the internet less, not more. I know the net neutrality debate is a little more complicated than that, but I look at most gov't regulation as a sort of 'gateway drug'. Once they start to regulate something, they can't stop. Heck, if they regulate it enough, they want to start taxing it too.
Could you imagine being the graphic artist tasked with having to come up with cover art for some of these games? What are those creatures on the cover of "PenPen Trilcelon"? Maybe the game designers where going for something emphasizing the intensity of war and the artist/marketing department was confused. I could see the same title being used with a "Dr. Strangelove" like theme.
You're assuming synthetic smells have a 'high-fidelity'. I don't know how accurate these smells really are. Do cheetos really taste cheesy? Can you make a soy or tofu burger taste or smell like beef? Sometimes knock-offs work, sometimes they don't. Incidentally, they already sell low-cost imitations of famous perfumes.
Yeah, studying EE would be my suggestion as well. At least, take some courses, complete a minor (if you can), or get a certificate. Heck, a grad school will look at your transcripts, so they really don't care if you get that second degree or not. A lot of math/statistics departments might be useful as well. Frankly some of the neatest research in AI is being done by EEs and statisticians. CS really doesn't give you the appropriate background. Also, most statisticians don't hack.
Stanford is pretty good too. At CMU they don't do interdisciplinary work, they just create a new department. Plus, Stanford's EE and Stats programs do good machine learning and information theory research as well.
In order to give a proper estimate, you need to be able to sit down with a piece of paper and break a 'job' down into it's atomic parts. While I'm being very vague about the word atomic, I'm doing it for a reason. It means different things in different contexts. In a year-long project, it may be an entire module, feature, or class. However, if you're breaking down your day's work, an atomic piece might be a particular method. Only when you've created a blueprint of your work and a priority list, can you effectively determine how long something will take. Oh, and multiply that number by 2 or 3.
Frankly, that's what your managers should be doing. Of course, they figure they can just slop an assignment in front of someone and ask how long it can take. Scotty says it will take a year, and he just can't 'give any more, cap'n'. When managers ask dumb questions, they get dumb answers.
See, there's really a much deeper political problem here. It looks like you customizing your existing app out the @$$ for your client and your boss is riding your behind to maintain undocumented code. This will deteriorate the code base! I'm guessing refactoring isn't a priority here. There's nothing wrong with doing this, provided you realize your code base will need to be thrown away if this trend continues. This practice has the additional effect of burning out your developers as well. Frankly, when you have little documentation low turnover is crucial. It also looks like there's a lot of bad software engineering going on here too. You're being asked to give estimates because managers can't give estimates. How can you even begin to follow the process I described above if you've got no requirement or design to begin with. My advice to you is guesstimate and CYOB. Also, it sounds like you're having to reword (or atleast clearify) a client's requirements. Once again, this sounds like a project manager's job. Sure, they can come to you for advice, but there's a point to where a manager needs to be taking some inititiative. Lastly, there needs to be some real management going on here. They need to cluster tasks. Often times multiple features are implemented in the same code, having multiple developers over a short period of time working on the same code with overlaping code enhancements is just not good appropriation of time. Of course, this all ties back into the lack of technical/project management from day one.
Oh, for your own sake, if this is what the whole company is like, look for a better job. Keep your feelers out, because you can't run a SaaS/ASP/software shop like this and make money.
What kind of car simulator was this? I remember (a long time ago), they used to use 'Test Drive' to run these types of tests. Do you think the physics engine was very accurate? I can drive a car, but I'm not very adept at playing video games. Ironically, games probably require more coordination than driving. Maturity is another issue all together. Assuming this game required more coordination than driving and the drivers were given very little alcohol, I think you make a claim the study was rigged.
If the telecommunication companies created a toll-booth internet, couldn't some other (even inferior) cheaper techonology clean up? Why do we have to 'use' their 'pipes' to send data. Sure we'd have to invent a new protocol (heck, I'd think the toll-booth thing would require that anyway), but couldn't we use satellite, wi-fi, and other forms of transmission?
Why do we constantly look for the "next big thing" when the "big thing" is simply experience?
When you do something once, it's science; when you do something twice, it's engineering. Experience is a funny thing. While I've learned a lot about software engineering, I still find differing problem domains to be challenging. Methodology doesn't just help me document a process, it allows me to wrap my head around a big COMPLEX problem. How do you solve a problem without solving it? You can't. Sure, it would be great if I could build an accounting system after I just built an accouting system, but life doesn't work that way. I'm also taking a shot in the dark that your average client doesn't want you to hack together a solution for them.
Which leads to the point that you either need a manager who has some understanding of the business process of software methodology or has a SENIOR developer (read 'architect') who he can rely on. A good manager in the software world needs to understand documentation. They really need quasi-technical skills. Sure, a CIO can leave the details to an architect, but somewhere a quasi-technical needs to be in the decision making loop to provide push-back so the company's "business needs" don't always overrun engineering sensibilities. (Yeah, I feel like a Dilbert comic strip today...)
It is my opinion that using an incremental deliverable approach with frequent customer meetings and executive power at any point in the project is the most successful strategy. The "rules" you have to adhere to are up to you and should be purely a case by case basis.
Isn't that the iterative methodology (e.g. the Unified Process and most Agile methods) in a nutshell?
Television isn't free -- they sell commercials Radio isn't free -- they also sell commercials Public education isn't free -- they use our tax money to support it Even PBS and NPR require donations.
OSS is actually free. The programmers may get a bit of extra experience that they can leverage into a job. Some even spin their software into a business. However, OSS is essentially free. Truthfully, most of the internet is free because it has deep roots in academia. Scientists and engineers (to a lesser extent) have always valued knowledge for its own sake. Is OSS as user friendly as commercial software. No. Is it as polished. No, but some of the most reliable products are OSS (Apache, Sendmail, the Linux OS itself).
This is way off topic, but I ran across an old post you made about Open Source Game Programming. I couldn't post a reply to that since it's so old. Normally I wouldn't do this, but I figured you might still be looking for a good RAD tool for game development. You might want to check out DarkBASIC. It's not OSS, but it isn't too expensive either.
Frankly, IIRC correctly your average retail store's receipt (especially store credit card applications) tend to have more personal information than most of what people do on a computer. I don't see retail stores shredding their trash. Dumpster diving and social engineering are probably the most numerous causes of identity theft today. (Yeah, all it takes is one really good hack to harm a lot of people.)
Slashdot Mirror
Are these deals with large manufacturers or "small" companies? On what forums do these deals take place? I'm not a gearhead, just interested in the economics of it all.
Yeah, when I read the headline "divide an image into discrete patterns", I was reminded of the 'Applied Probability Course' offered by ArsDigita.org. One of the lecturers appearantly did a PhD thesis on this topic. So, I was kind of impressed that some father would be interested in going to such lengths for his kid. (IIRC, there were only 5 lectures for the course, so you could certainly watch them if your interested in this subject.)
Tip: Most 'hard' problems have really nifty approximate solutions that are good enough for most applications (i.e. problems not in AI/automation/verification).
P.S. I think the algorithm using a 'histogram of top N colors' is pretty cool for a automated paint-by-number pattern generator. Seems like that could be a cool perl hack. Sure, you might have to throw a way some of the more 'complicated' images that have too many shades of a color (or other similar low probability scenarios). Maybe you could examine the difference in RGB values of colors in the pallete and examine the frequency in which these colors are used.
if it's any consolation, most logicians are so nitpicky, they actually annoy other types of mathematicians. It kind of makes you wonder what conversations with Godel was like?
Most restaurant owners aren't IT savvy. Especially if it involves exotic hardware, a non-windows OS, and sparsely documented software. Heck, most restaurant managers have problems using the windows machine in the back. From what I've seen (which admittedly isn't much) most restaurants hire a technician to maintain their POS setup. I hope the owner is paying the article poster. Frankly, he's going to need support no matter what.
If you want to write viruses and easy-to-use hacking tools -- yes. Look, Perl is the glue for *nix. However, thanks to Microsoft's object models, VB/VBA/Visual Studio Tools for Office is the probably more the Window hackers tool of choice that many would care to admit. Of course, how many respectable 'hackers' hack Windows?
A contract is only as good as your ability to enforce it in court. While legally you are correct in saying that a computer transaction can be made into a legally binding agreement, producing evidence that joe sixpack @ 111.111.111.111 is really who he says he is and collecting on that contract may be more difficult. Do I really want to bring a computer forensics expert or my sysadmin to court? Plus, tying a face, identification, and a 'physical presence' to a legal contract only adds to the legitimacy.
If you're bringing in $100,000+ per month in sales, you probably should get your own site. EBay is great for the 'small-time' or 'early-stage' seller. It's a robust engine that can handle heavy traffic and activity and is reasonably secure. However, if you've got sales as high as mentioned in the article, I can't imagine how EBay could provide an adequate solution for you. If you sell big ticket items, you probably want some binding contractual arrangement. On the other hand, if you have an incredibly high volume of sales, you probably want to better cater the shopper's experience and would have an E-commerce site of your own. Not to mention the fact that you'd need some internal system (even an Access database) to track all of your sales and shipping data. Maybe EBay's API allows you to do some of these things. I'd certainly be happy to hear about anyone experiences with it.
If it's any consolation, the matrix violates the laws of thermodynamics. Plus, the human body is a grossly inefficent battery. Mr. Sun is what helps the little 'open system' that is Earth go. In other words, the robots would probably kill us and use us a fertilizer instead of an energy source. So don't worry!
Let's face it. Public schools are large lumbering institutions. I learned very little there. Sure, the governor's school and advanced courses are usually pretty good, but the average public school education is filler at best. The modern day educator is afraid of legal suits, sexual harrassment charges, excessively disrespectful behavior by students (you know it's true), and a largely hostile collection of parents. Schools respond by imposing draconian rules and excessive punishment for relatively petty offenses. It's truly an administrative mess. Why congress isn't debating THIS 24/7 is truly mystifying. Unless I'm dead broke, I don't plan on sending my kid to a public school.
Ummm, alaskans.
The cold weather up there does strange things to them. Haven't you ever seen Northern Exposure?
If anything, I want these legislators regulating the internet less, not more. I know the net neutrality debate is a little more complicated than that, but I look at most gov't regulation as a sort of 'gateway drug'. Once they start to regulate something, they can't stop. Heck, if they regulate it enough, they want to start taxing it too.
Could you imagine being the graphic artist tasked with having to come up with cover art for some of these games? What are those creatures on the cover of "PenPen Trilcelon"? Maybe the game designers where going for something emphasizing the intensity of war and the artist/marketing department was confused. I could see the same title being used with a "Dr. Strangelove" like theme.
You're assuming synthetic smells have a 'high-fidelity'. I don't know how accurate these smells really are. Do cheetos really taste cheesy? Can you make a soy or tofu burger taste or smell like beef? Sometimes knock-offs work, sometimes they don't. Incidentally, they already sell low-cost imitations of famous perfumes.
Yeah, studying EE would be my suggestion as well. At least, take some courses, complete a minor (if you can), or get a certificate. Heck, a grad school will look at your transcripts, so they really don't care if you get that second degree or not. A lot of math/statistics departments might be useful as well. Frankly some of the neatest research in AI is being done by EEs and statisticians. CS really doesn't give you the appropriate background. Also, most statisticians don't hack.
Although, computer vision and voice recognition are popular areas of AI research. Certainly, the former is deeply routed in robotics.
Stanford is pretty good too. At CMU they don't do interdisciplinary work, they just create a new department. Plus, Stanford's EE and Stats programs do good machine learning and information theory research as well.
In order to give a proper estimate, you need to be able to sit down with a piece of paper and break a 'job' down into it's atomic parts. While I'm being very vague about the word atomic, I'm doing it for a reason. It means different things in different contexts. In a year-long project, it may be an entire module, feature, or class. However, if you're breaking down your day's work, an atomic piece might be a particular method. Only when you've created a blueprint of your work and a priority list, can you effectively determine how long something will take. Oh, and multiply that number by 2 or 3.
Frankly, that's what your managers should be doing. Of course, they figure they can just slop an assignment in front of someone and ask how long it can take. Scotty says it will take a year, and he just can't 'give any more, cap'n'. When managers ask dumb questions, they get dumb answers.
See, there's really a much deeper political problem here. It looks like you customizing your existing app out the @$$ for your client and your boss is riding your behind to maintain undocumented code. This will deteriorate the code base! I'm guessing refactoring isn't a priority here. There's nothing wrong with doing this, provided you realize your code base will need to be thrown away if this trend continues. This practice has the additional effect of burning out your developers as well. Frankly, when you have little documentation low turnover is crucial. It also looks like there's a lot of bad software engineering going on here too. You're being asked to give estimates because managers can't give estimates. How can you even begin to follow the process I described above if you've got no requirement or design to begin with. My advice to you is guesstimate and CYOB. Also, it sounds like you're having to reword (or atleast clearify) a client's requirements. Once again, this sounds like a project manager's job. Sure, they can come to you for advice, but there's a point to where a manager needs to be taking some inititiative. Lastly, there needs to be some real management going on here. They need to cluster tasks. Often times multiple features are implemented in the same code, having multiple developers over a short period of time working on the same code with overlaping code enhancements is just not good appropriation of time. Of course, this all ties back into the lack of technical/project management from day one.
Oh, for your own sake, if this is what the whole company is like, look for a better job. Keep your feelers out, because you can't run a SaaS/ASP/software shop like this and make money.
What kind of car simulator was this? I remember (a long time ago), they used to use 'Test Drive' to run these types of tests. Do you think the physics engine was very accurate? I can drive a car, but I'm not very adept at playing video games. Ironically, games probably require more coordination than driving. Maturity is another issue all together. Assuming this game required more coordination than driving and the drivers were given very little alcohol, I think you make a claim the study was rigged.
If the telecommunication companies created a toll-booth internet, couldn't some other (even inferior) cheaper techonology clean up? Why do we have to 'use' their 'pipes' to send data. Sure we'd have to invent a new protocol (heck, I'd think the toll-booth thing would require that anyway), but couldn't we use satellite, wi-fi, and other forms of transmission?
When you do something once, it's science; when you do something twice, it's engineering. Experience is a funny thing. While I've learned a lot about software engineering, I still find differing problem domains to be challenging. Methodology doesn't just help me document a process, it allows me to wrap my head around a big COMPLEX problem. How do you solve a problem without solving it? You can't. Sure, it would be great if I could build an accounting system after I just built an accouting system, but life doesn't work that way. I'm also taking a shot in the dark that your average client doesn't want you to hack together a solution for them.
Which leads to the point that you either need a manager who has some understanding of the business process of software methodology or has a SENIOR developer (read 'architect') who he can rely on. A good manager in the software world needs to understand documentation. They really need quasi-technical skills. Sure, a CIO can leave the details to an architect, but somewhere a quasi-technical needs to be in the decision making loop to provide push-back so the company's "business needs" don't always overrun engineering sensibilities. (Yeah, I feel like a Dilbert comic strip today...)
Isn't that the iterative methodology (e.g. the Unified Process and most Agile methods) in a nutshell?
Television isn't free -- they sell commercials
Radio isn't free -- they also sell commercials
Public education isn't free -- they use our tax money to support it
Even PBS and NPR require donations.
OSS is actually free. The programmers may get a bit of extra experience that they can leverage into a job. Some even spin their software into a business. However, OSS is essentially free. Truthfully, most of the internet is free because it has deep roots in academia. Scientists and engineers (to a lesser extent) have always valued knowledge for its own sake. Is OSS as user friendly as commercial software. No. Is it as polished. No, but some of the most reliable products are OSS (Apache, Sendmail, the Linux OS itself).
This is way off topic, but I ran across an old post you made about Open Source Game Programming. I couldn't post a reply to that since it's so old. Normally I wouldn't do this, but I figured you might still be looking for a good RAD tool for game development. You might want to check out DarkBASIC. It's not OSS, but it isn't too expensive either.
Frankly, IIRC correctly your average retail store's receipt (especially store credit card applications) tend to have more personal information than most of what people do on a computer. I don't see retail stores shredding their trash. Dumpster diving and social engineering are probably the most numerous causes of identity theft today. (Yeah, all it takes is one really good hack to harm a lot of people.)