At the various conferences and other tech events I go to, I've met many Sun and Microsoft employees. One thing that really strikes me is that I've yet to meet a Sun employee younger than about 35, but I've also never met a Microsoft employee (other than an executive) over 35. I think this creates problems for both companies.
I'll download it as soon as there's something better than a CAM out there:-)
While this was intended to be funny, there is some seriousness here. If I were Moore, and my goal was widest possible distribution, not most money made, putting it up for P2P download is a great idea. However, having gone to the trouble to shoot and edit the movie with high enough quality for movie projection, I'd want the highest possible quality to be downloadable. So, if he were really interested, he'd upload a copy from the original sources.
Or, maybe Moore sees P2P as an advertising medium to drive sales of movie tickets? In which case, you'd probably want a low-quality CAM capture to be floating around the net.
What's the best way to read online texts? There are a bunch of PG texts I might like to read, but reading them in a web browser, as a big text file gets tiring after ten minutes or so. I'm not sure why I can read a book for hours, but the screen for minutes, but there you have it. I don't think that HTML will help this problem -- does anyone have recommendations for better ways to read these files?
Many of us have been conditioned to think that both standards and innovation are good things. And the latter is an overused word that Microsoft marketing has forced into the memestream. But really, standards tend to stifle innovation.
Perhaps the most related, useful, but most underappreciated related discipline is statistics. Of all the non-CS classes I took, stats is the most relevant to my day-to-day life. For example, doing analysis of performance and tuning software system, I often see people use bogus statistical analysis, and making mistakes based on those results. Even if your curriculum doesn't require it, I would highly recommend taking a stats class or two.
In the early days of the patent office, they used to require a working model. I think they should return to this requirement. In my experience, ideas themselves, which can be patented, aren't nearly as valuable as the implementation of those ideas. This would probably also cut down on a lot of bogus patent submissions.
These best and brightest also create the foundations of your future product lines, patent licensing income, etc etc.
If this were true, Bell Labs wouldn't be shutting down. The truth is, none of the world-class pure research labs (Xerox PARC, Bell Labs, TJ Watson, etc.) do a good job of helping their parent companies in the long run. How much money did Bell Labs directly make on the transistor? The Laser? The C programming Language? C++? Unix succeeded in large parts despite the efforts of USL. Look at how successful Xerox PARC made Apple, but not Xerox. Very few companies now do much pure research for this very reason.
Many other companies (and society in general) tend to benefit from the breakthroughs in pure research from these labs, but the orginal companies usually have the biggest difficulties managing this.
IBM has a less well known Java VM for embedded systems called j9. This was developed in a clean-room way. If IBM wants an Open Source, commercial quality VM, there's nothing stopping them from opening this one.
Start with Cryptonomicon, polish it up with "hipper" characters, speed it up, dumb it down, speed it up again, remove Stephenson's on-an-on tendencies, and you've got Dan Brown's books.
The NYT ran an article about this a while back -- they interviewed people who didn't want to buy things from unsolicited phone calls, but ended up doing so anyways, because they "felt sorry" for the person making the phone call. One woman they interviewed bought 5 $1k replacement windows, which she later admitted, "probably wasn't a good idea". She also said that she would sign up for a do not call list, to prevent her from mercy purchases in the future.
Don't think that the telemarketters don't know their own business.
Ranting about the other side of the process
on
Joel Rants About Resumes
·
· Score: 4, Insightful
Joel's rant seems fair enough, but let's look at the other side of the process.
Again and again, I see job positions for which the applicant is asked to submit a resume via a textbox in a web form. Usually, no mention is given of what format is allowed (Plain ASCII? HTML? PDF? Tex?), so one pretty much has to assume least-common denominator, and submit in ASCII. Then, one has to pray to the line-width gods that the end product (printed out? online?) will not look too horrible compared to what you just put in.
So, for those on the other side of the table, can you please implement a simple web-file-upload protocol, and tell us what format you like?
memory protection adds overhead and can affect real-time performance
This is the conventional wisdom, and in my experience, this particular nugget causes more embedded and real time software projects to fail than any other.
First off, on a modern PowerPC processor, memory protection (that is, without virtual memory support) can be implemented very cheaply. If you can do it just with the IBAT/DBAT registers, it should be a constant-time overhead, which is good enough for hard-real time. Oddly enough, I can't find a single reference on the net that measures the cost of memory protection alone on a modern CPU. Anyone? Anyone?
Secondly, though the rover certainly may have some software components that have hard-real time requirements, that doesn't mean that every single line of code does. Typically, less than 1 percent of the code in a real time system is hard real time. In that case, you can run the real-time code in ISRs, or perhaps in a dual-mode system, like RT-Linux, or in high-priority kernel threads (as with QNX). In any of these situations, you can run all the rest of the code in protected memory space.
If I was sending an embedded control computer to another planet, I would have chosen an OS with memory protection, not VxWorks. VxWorks is like DOS, and early versions of Windows, where one pointer problem in one task can corrupt the whole system. Sure, we don't know that's the problem now, but it would be nice to know for sure that it wasn't.
Last time I looked into this, which was several years ago, there were no known different strings which had the same MD5 hash. I thought this was remarkable. Are there any known ones today?
Gee, Rocky, that M&A trick never works
on
EMC To Acquire VMware
·
· Score: 4, Interesting
I predict the merger will fail, horribly. Of course, that's an easy bet, given the history of most corporate merger and acquistion activity. Seems like 90% of them fail -- Wind River buying BSDi, Compaq & DEC, AOL & Netscape. And of those that "succeed", seems like the success isn't any better than what they would have been anyway -- e.g. Microsoft & Hotmail. Can anyone name an acquisition story that's been a huge success?
Keep in mind that these "Java phones" are running J2ME of some flavor or another, which is about as close to the desktop Java you are familiar with as Javascript is.
Re:Good database design is still essential
on
Bitter EJB
·
· Score: 1
What about OO-Mapping tools like Hibernate or Tophat (an Oracle product)?
Same problem. No matter what the marketing will tell you, without a solid database design under the hood, you are hosed.
Good database design is still essential
on
Bitter EJB
·
· Score: 4, Interesting
The big lie with EJB, and all the OO RDBMS mapping and access tools, is that you don't need to learn anything about relational databases, modeling or theory in order to use them successfully. For some reasons, most developers are terrified that becoming a database expert means becoming a tie-wearing, COBOL-writing "business programmer". Perhaps this is because database theory and application is taught horribly in most Universities.
Every failed EJB based system (and many failed non-EJB systems) that I've seen had poor database schema design at their heart. In almost every case, engineers familiar with OO design, didn't really bother to learn anything about RDBMS, perhaps because they felt them to be old-fashioned, and tried to ignore the database as much as possible, and put too much logic outside of the database.
There's a lot to EJBs, sure. But you first have to learn the basics of good database design, for without that, you are hosed.
I've got bad news for the marketting people out there -- they waste a lot more than half their marketing budget. I bet 90% of the advertisements I see are completely useless to me.
The only things that weren't in CVS were gcc, gmake, bison/flex.
My hard-core version control friends say that everything should be under version control in an ideal environment, including compiler and host operating system! I think that's a bit extreme, but I like to use the -V version option to gcc in my makefiles to at least document which gcc version the code requires.
I'm not asking this to be smarmy, I'm really curious, and I hope that I can get an answer, even though this reply is late in the thread. Every time I've seen software escrow (from both sides), it was a last-minute thing, and not really thought through well. On the developer side, every quarter, the lowest developer on the totem pole was asked to tar up a source tree, put it on tape, and FedEx it to the escrow company. Was that tree buildable? Who knows? Did it correspond to any version we shipped? Doubt it. Did we ship full build instructions? Nope.
For software escrow to really work, the escrow company has to be in between the developer and the end user. Whenever the developer wants to ship a new release, the escrow company should check it out of the source code control, build it themselves, and they should ship the binaries to the customers. This is the only way to ensure that the binaries the customers run are fully covered by escrow. But no one does that -- it is too expensive.
For a lot of these advisories, you can plug the hole at the firewall, or maybe the mail server. Do you really need to allow MS messenger service to be running outside your LAN? Sure, it is a good idea to quickly patch the systems, but it may take days to get them all patched. Fixing the problem outside the Windows boxes can be done within minutes of reading the advisory, depending on where the problem is.
Slashdot Mirror
At the various conferences and other tech events I go to, I've met many Sun and Microsoft employees. One thing that really strikes me is that I've yet to meet a Sun employee younger than about 35, but I've also never met a Microsoft employee (other than an executive) over 35. I think this creates problems for both companies.
While this was intended to be funny, there is some seriousness here. If I were Moore, and my goal was widest possible distribution, not most money made, putting it up for P2P download is a great idea. However, having gone to the trouble to shoot and edit the movie with high enough quality for movie projection, I'd want the highest possible quality to be downloadable. So, if he were really interested, he'd upload a copy from the original sources.
Or, maybe Moore sees P2P as an advertising medium to drive sales of movie tickets? In which case, you'd probably want a low-quality CAM capture to be floating around the net.
What's the best way to read online texts? There are a bunch of PG texts I might like to read, but reading them in a web browser, as a big text file gets tiring after ten minutes or so. I'm not sure why I can read a book for hours, but the screen for minutes, but there you have it. I don't think that HTML will help this problem -- does anyone have recommendations for better ways to read these files?
Many of us have been conditioned to think that both standards and innovation are good things. And the latter is an overused word that Microsoft marketing has forced into the memestream. But really, standards tend to stifle innovation.
Maybe these says more about the readers than the computer program?
And how, exactly, does a static compiler prove that a certain code sequence correctly computes a square root?
Perhaps the most related, useful, but most underappreciated related discipline is statistics. Of all the non-CS classes I took, stats is the most relevant to my day-to-day life. For example, doing analysis of performance and tuning software system, I often see people use bogus statistical analysis, and making mistakes based on those results. Even if your curriculum doesn't require it, I would highly recommend taking a stats class or two.
If you can't build it, how do you know it is such a good idea?
In the early days of the patent office, they used to require a working model. I think they should return to this requirement. In my experience, ideas themselves, which can be patented, aren't nearly as valuable as the implementation of those ideas. This would probably also cut down on a lot of bogus patent submissions.
If this were true, Bell Labs wouldn't be shutting down. The truth is, none of the world-class pure research labs (Xerox PARC, Bell Labs, TJ Watson, etc.) do a good job of helping their parent companies in the long run. How much money did Bell Labs directly make on the transistor? The Laser? The C programming Language? C++? Unix succeeded in large parts despite the efforts of USL. Look at how successful Xerox PARC made Apple, but not Xerox. Very few companies now do much pure research for this very reason.
Many other companies (and society in general) tend to benefit from the breakthroughs in pure research from these labs, but the orginal companies usually have the biggest difficulties managing this.
IBM has a less well known Java VM for embedded systems called j9. This was developed in a clean-room way. If IBM wants an Open Source, commercial quality VM, there's nothing stopping them from opening this one.
Start with Cryptonomicon, polish it up with "hipper" characters, speed it up, dumb it down, speed it up again, remove Stephenson's on-an-on tendencies, and you've got Dan Brown's books.
Don't think that the telemarketters don't know their own business.
Again and again, I see job positions for which the applicant is asked to submit a resume via a textbox in a web form. Usually, no mention is given of what format is allowed (Plain ASCII? HTML? PDF? Tex?), so one pretty much has to assume least-common denominator, and submit in ASCII. Then, one has to pray to the line-width gods that the end product (printed out? online?) will not look too horrible compared to what you just put in.
So, for those on the other side of the table, can you please implement a simple web-file-upload protocol, and tell us what format you like?
This is the conventional wisdom, and in my experience, this particular nugget causes more embedded and real time software projects to fail than any other.
First off, on a modern PowerPC processor, memory protection (that is, without virtual memory support) can be implemented very cheaply. If you can do it just with the IBAT/DBAT registers, it should be a constant-time overhead, which is good enough for hard-real time. Oddly enough, I can't find a single reference on the net that measures the cost of memory protection alone on a modern CPU. Anyone? Anyone?
Secondly, though the rover certainly may have some software components that have hard-real time requirements, that doesn't mean that every single line of code does. Typically, less than 1 percent of the code in a real time system is hard real time. In that case, you can run the real-time code in ISRs, or perhaps in a dual-mode system, like RT-Linux, or in high-priority kernel threads (as with QNX). In any of these situations, you can run all the rest of the code in protected memory space.
If I was sending an embedded control computer to another planet, I would have chosen an OS with memory protection, not VxWorks. VxWorks is like DOS, and early versions of Windows, where one pointer problem in one task can corrupt the whole system. Sure, we don't know that's the problem now, but it would be nice to know for sure that it wasn't.
Last time I looked into this, which was several years ago, there were no known different strings which had the same MD5 hash. I thought this was remarkable. Are there any known ones today?
I predict the merger will fail, horribly. Of course, that's an easy bet, given the history of most corporate merger and acquistion activity. Seems like 90% of them fail -- Wind River buying BSDi, Compaq & DEC, AOL & Netscape. And of those that "succeed", seems like the success isn't any better than what they would have been anyway -- e.g. Microsoft & Hotmail. Can anyone name an acquisition story that's been a huge success?
Keep in mind that these "Java phones" are running J2ME of some flavor or another, which is about as close to the desktop Java you are familiar with as Javascript is.
Same problem. No matter what the marketing will tell you, without a solid database design under the hood, you are hosed.
Every failed EJB based system (and many failed non-EJB systems) that I've seen had poor database schema design at their heart. In almost every case, engineers familiar with OO design, didn't really bother to learn anything about RDBMS, perhaps because they felt them to be old-fashioned, and tried to ignore the database as much as possible, and put too much logic outside of the database.
There's a lot to EJBs, sure. But you first have to learn the basics of good database design, for without that, you are hosed.
I've got bad news for the marketting people out there -- they waste a lot more than half their marketing budget. I bet 90% of the advertisements I see are completely useless to me.
My hard-core version control friends say that everything should be under version control in an ideal environment, including compiler and host operating system! I think that's a bit extreme, but I like to use the -V version option to gcc in my makefiles to at least document which gcc version the code requires.
For software escrow to really work, the escrow company has to be in between the developer and the end user. Whenever the developer wants to ship a new release, the escrow company should check it out of the source code control, build it themselves, and they should ship the binaries to the customers. This is the only way to ensure that the binaries the customers run are fully covered by escrow. But no one does that -- it is too expensive.
For a lot of these advisories, you can plug the hole at the firewall, or maybe the mail server. Do you really need to allow MS messenger service to be running outside your LAN? Sure, it is a good idea to quickly patch the systems, but it may take days to get them all patched. Fixing the problem outside the Windows boxes can be done within minutes of reading the advisory, depending on where the problem is.