... here. Don't be misled by the title of the article ("Mac poses as much of challenge to Linux and Windows"). Most of the article discusses the new RedHat release, with comparisons to M$ and MacOS X.
One (warning: self-serving promo coming) is to outsource to a consulting firm. They'll charge you a fee--but at the end of the project they will go away. You don't have any overhead costs, you don't have any headcount, and you don't have costs for machines and toolsets that you no longer need.
Yes, and the downside is that you also don't have anyone who really understands the code once the contract shop "goes away." Unless the contract is a long one, then once the code is written, it will have to be brought in house and tested and supported by someone, who is going to require time (and by extension, money) to be brought up to speed on said code in the event that something breaks or enhancements are requested.
One other point: maybe your firm is different, but I've suffered through enough poorly-written consulting-firm code to realize that said consulting firms have hardly cornered the market on hiring good software engineers, nor do they necessarily have any better understanding of good software engineering practices than an in-house developer does.
Into Networks has developed technology similar to what Valve is doing, and has been selling it for a while. A few sites offer software streaming using Into's technology; the offerings at EB1, Disney, and RealArcade are powered this way. In fact, when EB1 rolled out their site a year or so ago, it got a story on Slashdot (though I can't find the link to it at the moment)
That statement caught my eye, too. All I can say is that maybe it's not true for the software that he's written, but he hasn't written every single line of software code in operation, and yes, some code is bad enough to be taken out and shot.
So, when should code be replaced (i.e. taken out and shot)? IMHO, it's when the overall architecture is so bad or the wrong language or toolset is used such that refactoring simply won't work. In a lot of projects, not nearly enough time is spent on software architecture and design, and not enough of the proper expertise is on hand to come up with a good design. As a result, the code is often quite ugly, not just because it's not commented, but because the chosen approach was flat-out wrong. That's when code should just be junked. Should it be scrapped one module at a time (i.e. refactoring)? Well, that would be safer in terms of product quality, but that approach has to be thought out even very carefully than the "blank sheet of paper" approach, and not many coders or managers have the patience to do that.
PGP encryption could use a nice high profile use case where its use...
I think that the best use case that would forward the cause of encryption in general (not just PGP in particular) will more likely come from a "not use" case, where perhaps sensitive, unencrypted corporate data (e.g. a contract in MS Word format, financial results, etc.) is either intercepted or copied from the sender or receiver computers. Right now, computer security simply is not taken seriously enough in most companies to warrant spending the money/time/etc. for PGP or any other encryption product.
That's interesting: how does OS X speed compare to OS 9.2 on either of those machines that you mentioned? The early reviews on OS X painted it as quite a slow beast, especially wrt the new Finder.
And point taken on the usability department. Even with KDE, Linux has a lot of room for improvement there. Peripheral support, too. I haven't found any LinuxPPC drivers for my Imation Superdisk.
SourceForge is a good place to start if you wanted to get involved in OSS development, but I would question whether such experience is truly valiable in the job market.
Having just gone through the interview cycle, I can say that most employers ask what languages and technologies you know and have worked with, period. Doesn't matter if you back it up with an OSS project or commercial work. And I have seen more than one job postings that mentioned OSS development as a "plus".
it really doesn't demonstrate an ability to work in a close knit team, to meet deadlines, to solve problems, enguage in project management activities (in the more treditional sense)
Huh? Meeting a constant stream of deadlines is what college is all about. Anyways, employers understand that a college grad won't have many of those skills anyways. That's why he/she starts off as a junior developer, and if said employer is smart, will pair them up with a senior engineer from whom the junior developer can learn.
What does any breed of PPC linux offer now that OS X doesn't?
Of the two, only Linux PPC will run decently well on a non-G4 Macintosh. Wasn't OS X optimized for G4 machines? When I was looking to install an OS in addition to OS 9.1 on my G3/350rev1, I looked at Apple's "OS X supported on" page, and saw my machine at the bottom of the list. Given the (then) performance problems that OS X was having, it made the decision to install Linux PPC easy.
Oh, yeah, the YDL CDs were $30. Mac OS X is >$100?
Clean, compact code will not sell any more copies...
No, but there's a second variable in the profit equation, and that is cost, which for software means not just development cost but support costs also. Depending on what the code actually does, it probably costs a lot less to support, fix, and add enhancements to that clean, compact, well thought-out code than it does to support a buggy, bloated alternative.
The big problem here is that when working on the product specifications and plans, very few managers bother to even consider issues related to support and maintenance, much less actually try to anticipate and/or account for how much that support and maintenance is going to cost the company over time.
Their biggest users could have been corporate, but at a couple hundred bucks a shot, most corporations had a hard time convincing themselves it was worth it on a large scale...
Good point, but I think that there's more to it than that. I know of companies that don't want their employees having encryption products available (and of a few that outright ban them as a matter of policy). While none of these outfits come right out and say so, I'd imagine that if employees start using encryption, companies would have a much more difficult time monitoring employee e-mails. Sad, but probably true.
This of course won't stop anti-ad software from simply downloading the ad and not putting it on screen, but most blocking mechanisms just don't download the ad anyways
FWIW, I think that the iCab browser for Macintosh is one of the few that I've seen that can be configured to not display images of a certain size on the page, just leaving a blank where the ad would have been. And, it comes pre-configured with a list of typical ad/banner sizes to screen out. I believe that it's a "download but don't display mechanism". Is there anything like this in Konqueror (I haven't been able to find it).
Now seriously, how much trouble is it to read around a huge ad in the middle of your page?
Depends on how annoying the ad is. There are two issues here: one is that the ad itself can takes up computing resources to display. Maybe those with 2GHz P4 machines don't mind this, but for those of us working with older processors, it unacceptably slows down the page display.
The second is just how distracting the ad itself is. I'm still trying to train that "mental filtering" of which you speak to screen out images of fast-moving monkeys or flashing blue and white images that I've seen that distract from what I'm trying to read. Some of these ads have gotten so ridiculous that I can't read the page on which the ad is placed. And a concept that web marketers don't seem to grasp is that maybe people will remember a product backed by an annoying ad, but how many people will really buy that product if they negatively remember the ads for that product?
Here are some questions I have after reading the review:
Did the book's authors give one or more examples of what they consider a successful online/offline brand? Or are the supporting anecdotes given from lots of different companies? I don't see any examples from the book in the review itself.
What are the (book) author's qualifications for writing such a book? Are they themselves in charge of building an on-line brand for a bricks&mortar outfit? Are they web branding consultants?
Overall, why do they consider on-line branding any different from off-line branding? Isn't the general idea the same for either: "what does your company make, and why should we buy from you vs. your competitor?"
One problem is that none of the e-book devices on the market today are really well-suited for the end users to actually read an e-book:
Regular PCs aren't portable enough;
Notebook PCs are far more cumbersome than a real book, and really are the wrong form factor.
Handheld PCs have way too small of a display for reading anything more than a short paragraph.
The specialized e-book display devices are expensive compared to a real book.
The only way that I see e-books catching on is if lightweight, "tablet" PCs (something along the dimensions of 8"x11") ever get made and catch on.
This isn't to say that the e-book idea is inself a bad one. There were a few novels that I've read where I come across a character about midway into the book, and say "Now, who was he again, and where was he mentioned earlier?", and having a hyperlink to the spot in the text where the character was introduced would have been useful.
The validity of a given patent is inversely proportional to the job titles of the "inventors". Exhibit A:
The McAfee.com inventors of this patent
include Chandrasekar Balasubramaniam, CTO & Vice President of Engineering; Babu Katchapalayam, Director of Engineering; Core Products; and Ravi Kannan, Director of Engineering, Infrastructure Services.
The author basically says in the first 10 paragraphs or so that the current Internet hardware/software infrastructure simply isn't capable of supporting some of the more bandwidth-intensive applications (video-on-demand, smart devices, web-browsing cell phones) that some business-people have dreamed that it could support, and that the so-called smart devices (like interactive TV) haven't really taken off yet either, because they are too unwieldy to use. OK so far, but nothing I haven't already heard or read elsewhere.
But, he makes a very clumsy transition and uses three of the final five paragraphs in the article shilling for Microsoft, describing how the white knights at Redmond were riding in to save the day with Passport and.NET. How M$ is going to contribute to the infrastructure side of the equation is left as an exercise to the reader's imagination, I suppose.
In short, I know economy.com is not geared to techies, but it's not a very good article, just the same.
Last time I checked, Sun isn't shipping their Solaris product with IE either.
Ya know, I was about to type that the last time I checked, Microsoft doesn't make a version of IE that runs on Solaris. But I checked, and what do you know: they actually do have claim to have a beta of IE for Solaris (and I think I saw a version for HP-UX, also. I don't have access to a Sun box to verify that it works, but it is interesting nonetheless.
Speaking to the Java point, as long as the PC makers are allowed to ship a JVM ("are allowed" implies "without paying a higher price for Windows to Micro$oft"), this probably won't be an issue.
According to the Register in this article from yesterday, the DOJ tried to get the case expedited for the purpose of getting said injunction, but was denied by the U.S. Court of Appeals for the D.C. Circuit.
It's a good idea, except it would take a bit of a mindset shift at AOL to make it happen. Much like M$, AOL has traditionally gone the route of making their client more and more "feature-rich" (i.e. bloated) with each release, thus making it more and more unusable on those older PCs. I don't know what the minimum system requirements currently are to run the AOL client (and AOL's website doesn't specify it, either), but I'd be surprised that those older PCs could run AOL's current client very well, even if they substituted Windows for Linux.
Huh? mac.com isn't an ISP. It's part of iTools -- a collection of services (10Gb storage, e-mail address, other stuff) that Apple provides for free for buying a Mac.
You're probably thinking of Earthlink. They have a relationship with Apple similiar to the one that AOL is proposing with the PC makers. I don't think Earthlink has anything to do with mac.com.
DEC (in the "good ol' days") used to offer a couple of similar programs: GEEP (Graduate Engineering Education Program) and GEM (Graduate Education in Manufacturing). That benefit was really good: full tuition and full salary, while you attend grad school full-time. The only hitch was it was only offered to select employees, and the application process was fairly grueling, IIRC.
The problem with these programs was that a number of employees left almost right after getting the degrees that they earned entirely on company time, and for which the company got little -- if any -- payback. I remember one colleague who told me that she was planning to leave once she got her Master's (which was funded through this program), because she felt she could get more money once she left, and didn't feel she owed anything to the company.
Now, that was one of the more blatant examples of abuse that I saw, but I heard of more cases where graduates of these programs simply got their old jobs back once they had finished, rather than be given something more advanced and more in line with what they supposedly learned while in school. And that was just stupid on the part of their managers. In those cases, the employees simply got bored with their work after a short time and left. And in those cases, I don't fault those employees, since they were theoretically qualified for better work, which they weren't going to get.
What people don't realize is that company-sponsered education carries some "obligations" (for lack of a better word) for both sides. The employee has to at least try to make use of that training to help the employer in some way, and the manager should (by virtue of spending company money for the training) see that their employees are in a position to use what they've learned. Otherwise, it's just a waste of money and time for everyone involved, and over time won't be offered at all.
...perhaps its time to look at another line of work.
Frankly, that's pretty harsh. I don't think of a training budget as a way to "pay you to learn to do your job". That implies that the training budget should be used for remedial training, which I agree is a waste of $$$. However, it's to a company's benefit to offer training in tools and techniques that would make their employees more productive. And there has to be some room for experimentation in that regard as well. Not every tool or technique that's introduced will help the bottom line, but some will, and it's worth it to spend the time doing some research to find out.
As I said in another post, more and more companies are resorting to hiring consultants and outsourcing to fulfill software development projects that use new technologies, to save time and cost in training the internal employees. And in my experience:
the price of these consultants/outsourcing deals is much higher than the cost of training the internal employees;
the quality of work produced by the consultants isn't guaranteed to be any better than that which the employees could produce;
the project is usually thrown "over the wall" back to the company employees to support and maintain, anyways.
How much training do I think should be provided? I think a reasonable amount is three or four 1-week corporate training seminars per year or two college courses (which financially usually work out about the same).
In my recent experience, however, companies have gotten pretty stingy when it comes to offering training. The preferred solution for the past few years at different places at which I've worked -- especially when a relatively new technology is involved in the project -- is to either hire consultants or outsource a project altogether rather than provide training to then internal employees.
One bit of advice: a major issue to consider when training is available is getting practical experience with the technology in which the course is being taken. Taking a five-day corporate course in any technology isn't going to make anyone an expert. My advice for those who want to learn a new language or tool is to make sure there's a project on which you can apply and experiment with your newfound skills. Otherwise, the new skills will likely fade away very quickly, and also certainly won't help your resume at all.
Slashdot Mirror
I wonder if anyone inside of AOL has run the numbers to figure out
and compared that to the amount of revenue that they get from selling out their customers.
... here. Don't be misled by the title of the article ("Mac poses as much of challenge to Linux and Windows"). Most of the article discusses the new RedHat release, with comparisons to M$ and MacOS X.
One (warning: self-serving promo coming) is to outsource to a consulting firm. They'll charge you a fee--but at the end of the project they will go away. You don't have any overhead costs, you don't have any headcount, and you don't have costs for machines and toolsets that you no longer need.
Yes, and the downside is that you also don't have anyone who really understands the code once the contract shop "goes away." Unless the contract is a long one, then once the code is written, it will have to be brought in house and tested and supported by someone, who is going to require time (and by extension, money) to be brought up to speed on said code in the event that something breaks or enhancements are requested.
One other point: maybe your firm is different, but I've suffered through enough poorly-written consulting-firm code to realize that said consulting firms have hardly cornered the market on hiring good software engineers, nor do they necessarily have any better understanding of good software engineering practices than an in-house developer does.
Into Networks has developed technology similar to what Valve is doing, and has been selling it for a while. A few sites offer software streaming using Into's technology; the offerings at EB1, Disney, and RealArcade are powered this way. In fact, when EB1 rolled out their site a year or so ago, it got a story on Slashdot (though I can't find the link to it at the moment)
That statement caught my eye, too. All I can say is that maybe it's not true for the software that he's written, but he hasn't written every single line of software code in operation, and yes, some code is bad enough to be taken out and shot.
So, when should code be replaced (i.e. taken out and shot)? IMHO, it's when the overall architecture is so bad or the wrong language or toolset is used such that refactoring simply won't work. In a lot of projects, not nearly enough time is spent on software architecture and design, and not enough of the proper expertise is on hand to come up with a good design. As a result, the code is often quite ugly, not just because it's not commented, but because the chosen approach was flat-out wrong. That's when code should just be junked. Should it be scrapped one module at a time (i.e. refactoring)? Well, that would be safer in terms of product quality, but that approach has to be thought out even very carefully than the "blank sheet of paper" approach, and not many coders or managers have the patience to do that.
PGP encryption could use a nice high profile use case where its use...
I think that the best use case that would forward the cause of encryption in general (not just PGP in particular) will more likely come from a "not use" case, where perhaps sensitive, unencrypted corporate data (e.g. a contract in MS Word format, financial results, etc.) is either intercepted or copied from the sender or receiver computers. Right now, computer security simply is not taken seriously enough in most companies to warrant spending the money/time/etc. for PGP or any other encryption product.
That's interesting: how does OS X speed compare to OS 9.2 on either of those machines that you mentioned? The early reviews on OS X painted it as quite a slow beast, especially wrt the new Finder.
And point taken on the usability department. Even with KDE, Linux has a lot of room for improvement there. Peripheral support, too. I haven't found any LinuxPPC drivers for my Imation Superdisk.
SourceForge is a good place to start if you wanted to get involved in OSS development, but I would question whether such experience is truly valiable in the job market.
Having just gone through the interview cycle, I can say that most employers ask what languages and technologies you know and have worked with, period. Doesn't matter if you back it up with an OSS project or commercial work. And I have seen more than one job postings that mentioned OSS development as a "plus".
it really doesn't demonstrate an ability to work in a close knit team, to meet deadlines, to solve problems, enguage in project management activities (in the more treditional sense)
Huh? Meeting a constant stream of deadlines is what college is all about. Anyways, employers understand that a college grad won't have many of those skills anyways. That's why he/she starts off as a junior developer, and if said employer is smart, will pair them up with a senior engineer from whom the junior developer can learn.
What does any breed of PPC linux offer now that OS X doesn't?
Of the two, only Linux PPC will run decently well on a non-G4 Macintosh. Wasn't OS X optimized for G4 machines? When I was looking to install an OS in addition to OS 9.1 on my G3/350rev1, I looked at Apple's "OS X supported on" page, and saw my machine at the bottom of the list. Given the (then) performance problems that OS X was having, it made the decision to install Linux PPC easy.
Oh, yeah, the YDL CDs were $30. Mac OS X is >$100?
Clean, compact code will not sell any more copies...
No, but there's a second variable in the profit equation, and that is cost, which for software means not just development cost but support costs also. Depending on what the code actually does, it probably costs a lot less to support, fix, and add enhancements to that clean, compact, well thought-out code than it does to support a buggy, bloated alternative.
The big problem here is that when working on the product specifications and plans, very few managers bother to even consider issues related to support and maintenance, much less actually try to anticipate and/or account for how much that support and maintenance is going to cost the company over time.
Their biggest users could have been corporate, but at a couple hundred bucks a shot, most corporations had a hard time convincing themselves it was worth it on a large scale...
Good point, but I think that there's more to it than that. I know of companies that don't want their employees having encryption products available (and of a few that outright ban them as a matter of policy). While none of these outfits come right out and say so, I'd imagine that if employees start using encryption, companies would have a much more difficult time monitoring employee e-mails. Sad, but probably true.
This of course won't stop anti-ad software from simply downloading the ad and not putting it on screen, but most blocking mechanisms just don't download the ad anyways
FWIW, I think that the iCab browser for Macintosh is one of the few that I've seen that can be configured to not display images of a certain size on the page, just leaving a blank where the ad would have been. And, it comes pre-configured with a list of typical ad/banner sizes to screen out. I believe that it's a "download but don't display mechanism". Is there anything like this in Konqueror (I haven't been able to find it).
Now seriously, how much trouble is it to read around a huge ad in the middle of your page?
Depends on how annoying the ad is. There are two issues here: one is that the ad itself can takes up computing resources to display. Maybe those with 2GHz P4 machines don't mind this, but for those of us working with older processors, it unacceptably slows down the page display.
The second is just how distracting the ad itself is. I'm still trying to train that "mental filtering" of which you speak to screen out images of fast-moving monkeys or flashing blue and white images that I've seen that distract from what I'm trying to read. Some of these ads have gotten so ridiculous that I can't read the page on which the ad is placed. And a concept that web marketers don't seem to grasp is that maybe people will remember a product backed by an annoying ad, but how many people will really buy that product if they negatively remember the ads for that product?
Here are some questions I have after reading the review:
Did the book's authors give one or more examples of what they consider a successful online/offline brand? Or are the supporting anecdotes given from lots of different companies? I don't see any examples from the book in the review itself.
What are the (book) author's qualifications for writing such a book? Are they themselves in charge of building an on-line brand for a bricks&mortar outfit? Are they web branding consultants?
Overall, why do they consider on-line branding any different from off-line branding? Isn't the general idea the same for either: "what does your company make, and why should we buy from you vs. your competitor?"
One problem is that none of the e-book devices on the market today are really well-suited for the end users to actually read an e-book:
The only way that I see e-books catching on is if lightweight, "tablet" PCs (something along the dimensions of 8"x11") ever get made and catch on.
This isn't to say that the e-book idea is inself a bad one. There were a few novels that I've read where I come across a character about midway into the book, and say "Now, who was he again, and where was he mentioned earlier?", and having a hyperlink to the spot in the text where the character was introduced would have been useful.
The validity of a given patent is inversely proportional to the job titles of the "inventors". Exhibit A:
The McAfee.com inventors of this patent include Chandrasekar Balasubramaniam, CTO & Vice President of Engineering; Babu Katchapalayam, Director of Engineering; Core Products; and Ravi Kannan, Director of Engineering, Infrastructure Services.
You make the call...
The author basically says in the first 10 paragraphs or so that the current Internet hardware/software infrastructure simply isn't capable of supporting some of the more bandwidth-intensive applications (video-on-demand, smart devices, web-browsing cell phones) that some business-people have dreamed that it could support, and that the so-called smart devices (like interactive TV) haven't really taken off yet either, because they are too unwieldy to use. OK so far, but nothing I haven't already heard or read elsewhere.
But, he makes a very clumsy transition and uses three of the final five paragraphs in the article shilling for Microsoft, describing how the white knights at Redmond were riding in to save the day with Passport and .NET. How M$ is going to contribute to the infrastructure side of the equation is left as an exercise to the reader's imagination, I suppose.
In short, I know economy.com is not geared to techies, but it's not a very good article, just the same.
Last time I checked, Sun isn't shipping their Solaris product with IE either.
Ya know, I was about to type that the last time I checked, Microsoft doesn't make a version of IE that runs on Solaris. But I checked, and what do you know: they actually do have claim to have a beta of IE for Solaris (and I think I saw a version for HP-UX, also. I don't have access to a Sun box to verify that it works, but it is interesting nonetheless.
Speaking to the Java point, as long as the PC makers are allowed to ship a JVM ("are allowed" implies "without paying a higher price for Windows to Micro$oft"), this probably won't be an issue.
According to the Register in this article from yesterday, the DOJ tried to get the case expedited for the purpose of getting said injunction, but was denied by the U.S. Court of Appeals for the D.C. Circuit.
It's a good idea, except it would take a bit of a mindset shift at AOL to make it happen. Much like M$, AOL has traditionally gone the route of making their client more and more "feature-rich" (i.e. bloated) with each release, thus making it more and more unusable on those older PCs. I don't know what the minimum system requirements currently are to run the AOL client (and AOL's website doesn't specify it, either), but I'd be surprised that those older PCs could run AOL's current client very well, even if they substituted Windows for Linux.
Huh? mac.com isn't an ISP. It's part of iTools -- a collection of services (10Gb storage, e-mail address, other stuff) that Apple provides for free for buying a Mac.
You're probably thinking of Earthlink. They have a relationship with Apple similiar to the one that AOL is proposing with the PC makers. I don't think Earthlink has anything to do with mac.com.
...can be found in this morning's Boston Globe (story link here)
DEC (in the "good ol' days") used to offer a couple of similar programs: GEEP (Graduate Engineering Education Program) and GEM (Graduate Education in Manufacturing). That benefit was really good: full tuition and full salary, while you attend grad school full-time. The only hitch was it was only offered to select employees, and the application process was fairly grueling, IIRC.
The problem with these programs was that a number of employees left almost right after getting the degrees that they earned entirely on company time, and for which the company got little -- if any -- payback. I remember one colleague who told me that she was planning to leave once she got her Master's (which was funded through this program), because she felt she could get more money once she left, and didn't feel she owed anything to the company.
Now, that was one of the more blatant examples of abuse that I saw, but I heard of more cases where graduates of these programs simply got their old jobs back once they had finished, rather than be given something more advanced and more in line with what they supposedly learned while in school. And that was just stupid on the part of their managers. In those cases, the employees simply got bored with their work after a short time and left. And in those cases, I don't fault those employees, since they were theoretically qualified for better work, which they weren't going to get.
What people don't realize is that company-sponsered education carries some "obligations" (for lack of a better word) for both sides. The employee has to at least try to make use of that training to help the employer in some way, and the manager should (by virtue of spending company money for the training) see that their employees are in a position to use what they've learned. Otherwise, it's just a waste of money and time for everyone involved, and over time won't be offered at all.
Frankly, that's pretty harsh. I don't think of a training budget as a way to "pay you to learn to do your job". That implies that the training budget should be used for remedial training, which I agree is a waste of $$$. However, it's to a company's benefit to offer training in tools and techniques that would make their employees more productive. And there has to be some room for experimentation in that regard as well. Not every tool or technique that's introduced will help the bottom line, but some will, and it's worth it to spend the time doing some research to find out.
As I said in another post, more and more companies are resorting to hiring consultants and outsourcing to fulfill software development projects that use new technologies, to save time and cost in training the internal employees. And in my experience:
How much training do I think should be provided? I think a reasonable amount is three or four 1-week corporate training seminars per year or two college courses (which financially usually work out about the same).
In my recent experience, however, companies have gotten pretty stingy when it comes to offering training. The preferred solution for the past few years at different places at which I've worked -- especially when a relatively new technology is involved in the project -- is to either hire consultants or outsource a project altogether rather than provide training to then internal employees.
One bit of advice: a major issue to consider when training is available is getting practical experience with the technology in which the course is being taken. Taking a five-day corporate course in any technology isn't going to make anyone an expert. My advice for those who want to learn a new language or tool is to make sure there's a project on which you can apply and experiment with your newfound skills. Otherwise, the new skills will likely fade away very quickly, and also certainly won't help your resume at all.
Didn't we just have this conversation a few weeks ago?