"rpm" is a package format and tool. Its equivalent in the Debian world is "dpkg".
"apt-get" is a tool for network updating of distributions. Its equivalent in the "rpm" world are "rpmfind", "urpmi", and a bunch of others.
In my experience, "apt-get" works quite well while "rpmfind" and "urpmi" don't. The reason for that isn't some technical deficiency of "rpm", it's one of user community: there isn't the same dedicated group of volunteers keeping network repositories for "rpm" up-to-date and available. rpmfind.net is a great effort, but it's a one-man operation.
So, should we all switch to Debian packages? I don't think so. I think the dependency management in "rpm" packages is better and more robust than that in "dpkg", and I believe "rpm" packages encourage third party contributions (outside a single monolithic development organization like Debian) more.
If there is demand, maybe people can volunteer and bring systems like "urpmi" or "rpmfind" up to speed. Several commercial vendors are also trying to fill the gap.
Of course, a more fundamental question to me is whether continuous updating is even desirable. In most corporate environments, stability is probably more important than getting the latest bug fixes (yes, even getting the latest security bug fixes).
I have maintained both systems with and without ACLs and with hundreds of users. Current UNIX access control is primitive and maybe something more powerful is desirable. But in my experience, ACLs just aren't the answer.
ACLs allow users to come up with quick workarounds to access control problems. It's easy for the administrator in the short run because people won't bother you about this or that. But you end up with an incomprehensible mess of access rights, and you end up with lots of questions of why something does or does not work.
The traditional UNIX system, instead, forces people to think ahead and define clearly what they are trying to accomplish. Then they define a group for that purpose. The end result is easier to understand and easier to analyze.
On balance, I'd rather not have ACLs. There is value in keeping things simple.
Many user interfaces have allowed the user to change not just appearance but also behavior on the fly. Apple's and Microsoft's user interfaces have been the exception with their static, hardcoded, cumbersome appearances and behaviors.
Furthermore, even if there didn't exist decades of prior art, this kind of "themability" is a natural result of using standard object-oriented design principles and using dynamic object-oriented runtimes: creating separate software components communicating via abstract interfaces for appearance, user interaction, and behavior is a natural evolution of library design for graphical user interfaces.
What this patent mainly demonstrates is incompetence and ignorance on the part of the Apple engineers that applied for it.
It's also part of another disturbing trend: companies like Apple and Microsoft have depressed the quality of software libraries and software engineering so much that things that used to be pretty simple and straightforward might now be considered major breakthrough inventions in their systems.
There isn't much one can do about the patent office. But what you can do is examine patents critically when people apply for jobs. A long list of patents is not necessarily a recommendation, in particular when a closer list at the patents suggests that the people who wrote them simply aren't familiar with key developments in their field.
No, it was already and old idea when the patent was filed. The reason why it isn't used more was because web standards had moved on and most people decided to use other techniques for authentication when applicable.
It was then, and continues to be, used when sending URLs that contain automatic login information through channels that don't support cookies, like mail and IM services.
Well, the law says that it needs to be obvious to "someone of ordinary skill in the art", not "someone of or below ordinary skill in the art". Clearly, you weren't skilled enough, so you don't count. If the thing is so obvious that beginner (or the MBAs and layers) can understand it, it's evidently so obvious that it deservers a patent. It's even more deserving of a patent if the people who are actually of ordinary skill have decided long ago that it's a bad idea in most applications and it is therefore not used much, or that it is so obvious that nobody ever bothered to "publish" it other than embedded in open source programs.
Now, as some lawyers who make millions on this stuff never tire to point out: "Kids, you really have got to stick to the letter of the law. And, besides: all this is for your own good, and don't forget to read all your claims and floss regularly. When you all grow up and become laywers yourself, make lots of money and get hot babes, and a cocain addiction, you'll thank me for it."
Actually, I have run USENET servers, and, yes, it can be a pain. But that's not relevant.
For archiving, all you need is a leaf node connection with no users. You don't even ever have to store incoming messages in a news hierarchy, you just send them off to your archival storage system as they come in.
Another choice might be to run a traditional news server, turn off article expiration, and keep the news hierarchy on a file system with a hierarchical storage manager (too heavy-handed for me).
Traditionally, postings were assumed to expire within a standard period (a few weeks). You could override that with headers, and people used to do that pretty frequently. DejaNews ignored those headers.
Nowadays, I think most peopled don't bother setting headers. But USENET is still a discussion medium. Just because a few companies decided at some point to archive the stuff doesn't mean that the user's presumption should change.
The DejaNews service and its scalability aren't at issue.
What is at issue is the archiving of USENET messages, and archiving them isn't difficult. If you can't figure it out, you can even buy commercial off-the-self solutions to deal with it.
Most likely, it would mean that Microsoft would not be permitted to continue to use the code in question and would have to prove that they have removed it from their sources.
This is actually not much different from what was happening with BSD and AT&T UNIX: BSD needed to be cleaned before getting released.
When you post to usenet, you're sending your work to whatever every archives are in place, and you know it.
USENET postings have expiration dates, which state to recipients how long the authors intends for them to be retained on public servers. Keeping postings any longer than that looks like a pretty clear violation of copyright. The argument that "you know when you post that..." is bogus. Music publishers also "know when they publish CDs" that their music will get copied, but that doesn't invalidate their copyright.
The fact is that DejaNews got away with this because they were big, and, hey, what could a random USENET poster do?
What deja.com is selling is not the rights to the posts, or the posts themselves, but the work that they put into archiving the posts, which is considerable.
That's nonsense. There is no "work" involved in archiving USENET postings. DejaNews didn't manually classify or edit articles. What you need for USENET archiving is storage and a simple script. The whole USENET system is designed for easy, reliable replication and archiving. Yes, the storage costs money and it is "work" to buy new CDs and disk drives. But a "software pirate" doesn't acquire a copyright to the stuff he is copying just because it is takes time for him to copy the stuff.
News articles are implicitly owned by the author, and they expire. DejaNews already has violated copyright by keeping the content longer than its expiration date. Redistributing the archives as a whole ist just not acceptable as far as I'm concerned. USENET is a discussion forum, not a long-term archive.
Google should become more compliant with copyright law, not less. But, then, the company blatantly copies and retains other content as well ("cached pages").
These kinds of copyright cases aren't jury cases. Furthermore, access through an ISP isn't a right, its based on a contractual business relationship between you and your ISP, which, if you read it, can be canceled at will by your ISP. You don't have to do anything wrong to have your account canceled. That's the legal side. Sorry, that's the way the cookie crumbles, at least in the US.
It should be pretty obvious to the person posting it whether they have the copyright, and if they post anyway, well, as I was saying... that's the ethical side.
Tempting as it may be, I think the Free Software Foundation has always regarded such restrictions as undesirable, and I believe the open source definition also excludes it.
The reasoning is that free software and open source software is about principles, and it's bad to compromise those principles for some possible short-term gain.
From a purely practical point of view, I think it would just be a bad precedent for people to write such specific restrictions into licenses. Enforcement is difficult, and you can't easily revoke the restrictions when the company changes (many companies have come around to open source).
In this particular case, I don't even see anything particularly serious to get upset about. The guy is going after people who post commercial software to newsgroups. That would be a worthwhile thing to do even from the point of view of keeping some junk off of USENET.
From an ethical point of view, if someone posts an obviously copyrighted commercial software package to USENET, I think he deserved to lose access to the ISP. Folks, do something productive instead: learn programming and drive Steinberg and Powell's other clients out of existence by writing better open source software; don't blindly post their stuff.
From a technical point of view, someone who wants to post potentially incriminating material and uses NNTP from their cable modem deserves what they get.
As for the transparent proxies detecting content, unless there is legislation restricting users to a handfull of well-known formats, in the presence of even simple cryptography or compression, that's a pointless exercise. OTOH, I see no problem with companies trying to detect copyrighted works by participating in P2P networks; of course, since P2P is moving towards tit-for-tat schemes, these participants better be prepared to put up some interesting content themselves in order to particpate:-)
The link doesn't work with Netscape 4.7* on Linux: it gets lots of JavaScript errors. Without JavaScript, you get a blank page. And if you don't have cookies enabled, you get a "Browser Error". The recommendation of law.com: "enable all cookies".
Aren't you glad to know that law.com has your privacy and consumer choice at heart? As long as you allow them to track you and use the latest version of Internet Explorer, you are fine.
I believe both Ximian and RedHat are planning on making money off subscription upgrade services, and that looks a little like the fox guarding the hen house to me. If the people responsible for creating a lot of the software distributions and infrastructure also sell the subscription necessary to deal with that stuff, what incentive is there to make future distributions and package management simpler and more reliable?
While I'm not saying it will go that far, the logical end point of this might be that you go to a web site, type in the bits and pieces of stuff you want, and have your raw disk image updated regularly. User-visible client-side package management would become irrelvant, you'd be completely dependent on the subscription, and you could forget about CD-ROMs or other disconnected installations.
Yes, some form of convenient download and dependency checking is needed. And both RedHat and Ximian probably have the best intentions. But I'm not convinced that this is the right overall route to go in the long run.
The same is true for RPM-based distributions: you get dependencies, automatic updates, channels, etc. The technical details are different and it's used less frequently with RPM but it's there if you want it.
In my experience, none of the automatic upgrades are entirely reliable. Debian is so dependent on automatic upgrades that packaging bugs get fixed more frequently, but problems occur even with Debian and can be very annoying. I see little reason to upgrade continuously. If it isn't broken, don't fix it.
Microsoft has been pilfering open source software (usually software with a BSD license) for years and shipping it with their systems. And, actually, I don't think this is even bad: it probably has helped make Microsoft a little more compatible with the rest of the world. Still, I would not be surprised if some GNU software had found its way into Windows as well.
As for the GPL license, I don't particularly like the policy behind it, and I think RMS's recent approach to using it is too heavy handed. But, then, I don't particularly like Microsoft's "we own your firstborn son" licenses either. The legal foundations on which they are based are similar, and if Allchin complains about the GPL, well, he might lead the way and get rid of Microsoft's heavy-handed licenses. See, what Allchin is really saying is that he is concerned when other people use copyright and software licensing laws in ways that interferes with Microsoft's business.
And he probably senses that economically open source software is a threat to his company because open source makes sense: it is finally a good mechanism to distribute software development costs equally, whereas all Microsoft can offer is having customers pay hundreds of dollars year after year for "improvements" they don't want in the first place.
Well, tough cookies, Allchin. Basic economic theory tells us that in an efficient market, Microsoft won't be making much in the way of profit. So, learn to live with it, and innovate a bit on the software side for a change.
XML--logical next step up from UNIX text files
on
Inside XML
·
· Score: 2
UNIX used to rely on tabular data a lot: that's what awk, grep, cut, paste, and other UNIX programs were designed to deal with. That was a great system, because it allowed you to deal with huge amounts of data on machines with very limited memory.
The problem with tabular data is that it doesn't let you represent a lot of information in a convenient form: people often do need and want hierarchical/tree-structured data. And tabular data in UNIX isn't self-describing. Configuration files, package descriptions, bibliographic citatinos, etc. all need fairly complex descriptions.
There are many ways of representing tree structured data. XML wouldn't be my favorite, but it is workable. And XML is getting a fairly complete set of tools for dealing with tree structured data: search, extraction, restructuring, etc.
With this, maybe the Linux community will pick up some of that old UNIX spirit again. Today, the habit seems to be that when anything needs to get done on Linux, someone writes a big Gnome or command line program in C, or, on a good day, writes a monolithic Perl program. For example, something like "rpmfind" should really just be a collection of a few command line tools: something like the Xerces tools for extracting information, gunzip to uncompress the data, and curl to retrieve information. The same is true for a lot of other Linux applications.
Oh, if you want to play around with XML, I found some of the Apache Xerces tools at http://xml.apache.org/ to be quite useful. They come in both Java and C++ flavors.
The Sharp machines have always been a good value and a nice form factor. I think what has held them back was their hotsyncing software and limitations on programmability.
PDP-11's ran UNIX in much less than 512k of total memory, including networking. Linux isn't quite as miserly as that, but you can probably fit a Linux kernel, a standard C library, X client libraries, and an X server into less than 3Mbytes of space, perfectly acceptable for PDA standards.
I think Mozilla, Gtk+, Qt, and the C++ libraries are so huge because people these days write for reuse and generality, not simplicity, maintainability, space, or efficiency. I have concluded it's pointless to argue that point. Most programmers educated in the last 10 years don't even believe that they are doing anything wrong when they do that.
Chess, Go, and card games don't have a narrative.
In fact, I have a hard time thinking of any classic (non-computer) game that is about narrative. It's just hard to get emotionally involved, something I'd consider a part of succesful narrative, with a bunch of little wooden blocks that you push around a board.
Maybe games based on good narrative can be developed eventually. But until someone figures out how, game designers should perhaps just focus on making good games. There is ample precedent for how to go about doing that. And it neither involves making characters more "life like" nor adding a lot of bogus story background to the game.
It is not in Microsoft's interest to separate hardware from software sales. Despite the wide availability of Linux, not paying for Windows and other Microsoft products on computers from Compaq, Gateway, and probably others is very difficult. Why would Microsoft want to give up the competitive advantage of per-unit licensing and bundling? If they can charge not just for every Windows machine but every Linux machine, why shouldn't they?
Probably, you'll get the worst of both worlds: you buy the PC with a one-year subscription, to be activated when you first plug in the machine. You won't be able to get the PCs without those pre-paid subscriptions.
Besides, the PC and Windows platforms are so messy that things have to be preinstalled; the "put in CD and turn on" isn't just undesirable from a business point of view for Microsoft and PC vendors, it is technically hard.
If hardware vendors went back to selling plain hardware without any software, that would be great for Linux and Windows-alternatives. It's not going to happen.
Whether it's 100000 genes or 30000 genes, it has always been clear that how those genes correspond to physical traits is at best a very complex proposition. After all, genes code for proteins, not for body parts. If that wasn't clear to Gould, maybe he should have paid more attention to molecular biology.
The reason why many people believe that genes correspond to morphologies is because, despite the fact that genes code for proteins, it is empirically justified in a certain sense: genetic alterations in specific genes have been observed frequently to correspond to specific morphologies. Of course, that can be explained because those are the only alterations that are easily observed. It is likely that many other mutations have such widespread effects that they just cause death during development. And molecular biology has given us ample examples for the fact that many genes have many functions throughout development (e.g., "combinatorial codes").
All this has been abundantly clear to many molecular biologists for decades, including Gould's colleagues at Harvard. Unfortunately, there seems to be a class of scientists that make a name for themselves by stating that everybody else in their field is plain dumb and then restating the obvious as if they were the first to think of it. Sorry, but as a far as I'm concerned, while Gould expresses things very well, he hasn't said anything of interest in many years.
We may laugh this off, but this kind of nonsense has a corrosive effect. Even if I tell you that I'm going to tell you a lie, a false statement I make to you has been shown to influence your thinking and judgement later. Multiply that effect by several hundred million and who knows what happens? How much funding is NASA going to lose because of the false impressions this kind of show creates? The only thing I can see doing about this sort of thing is to express your outrage. If you subscribe to cable, unsubscribe and let them know why. When it comes to the dangers of media, this is the kind of stuff politicians should worry about.
Slashdot Mirror
"apt-get" is a tool for network updating of distributions. Its equivalent in the "rpm" world are "rpmfind", "urpmi", and a bunch of others.
In my experience, "apt-get" works quite well while "rpmfind" and "urpmi" don't. The reason for that isn't some technical deficiency of "rpm", it's one of user community: there isn't the same dedicated group of volunteers keeping network repositories for "rpm" up-to-date and available. rpmfind.net is a great effort, but it's a one-man operation.
So, should we all switch to Debian packages? I don't think so. I think the dependency management in "rpm" packages is better and more robust than that in "dpkg", and I believe "rpm" packages encourage third party contributions (outside a single monolithic development organization like Debian) more.
If there is demand, maybe people can volunteer and bring systems like "urpmi" or "rpmfind" up to speed. Several commercial vendors are also trying to fill the gap.
Of course, a more fundamental question to me is whether continuous updating is even desirable. In most corporate environments, stability is probably more important than getting the latest bug fixes (yes, even getting the latest security bug fixes).
ACLs allow users to come up with quick workarounds to access control problems. It's easy for the administrator in the short run because people won't bother you about this or that. But you end up with an incomprehensible mess of access rights, and you end up with lots of questions of why something does or does not work.
The traditional UNIX system, instead, forces people to think ahead and define clearly what they are trying to accomplish. Then they define a group for that purpose. The end result is easier to understand and easier to analyze.
On balance, I'd rather not have ACLs. There is value in keeping things simple.
Furthermore, even if there didn't exist decades of prior art, this kind of "themability" is a natural result of using standard object-oriented design principles and using dynamic object-oriented runtimes: creating separate software components communicating via abstract interfaces for appearance, user interaction, and behavior is a natural evolution of library design for graphical user interfaces.
What this patent mainly demonstrates is incompetence and ignorance on the part of the Apple engineers that applied for it.
It's also part of another disturbing trend: companies like Apple and Microsoft have depressed the quality of software libraries and software engineering so much that things that used to be pretty simple and straightforward might now be considered major breakthrough inventions in their systems.
There isn't much one can do about the patent office. But what you can do is examine patents critically when people apply for jobs. A long list of patents is not necessarily a recommendation, in particular when a closer list at the patents suggests that the people who wrote them simply aren't familiar with key developments in their field.
It was then, and continues to be, used when sending URLs that contain automatic login information through channels that don't support cookies, like mail and IM services.
Now, as some lawyers who make millions on this stuff never tire to point out: "Kids, you really have got to stick to the letter of the law. And, besides: all this is for your own good, and don't forget to read all your claims and floss regularly. When you all grow up and become laywers yourself, make lots of money and get hot babes, and a cocain addiction, you'll thank me for it."
(Smiley for the humor impaired: :-)
For archiving, all you need is a leaf node connection with no users. You don't even ever have to store incoming messages in a news hierarchy, you just send them off to your archival storage system as they come in.
Another choice might be to run a traditional news server, turn off article expiration, and keep the news hierarchy on a file system with a hierarchical storage manager (too heavy-handed for me).
Nowadays, I think most peopled don't bother setting headers. But USENET is still a discussion medium. Just because a few companies decided at some point to archive the stuff doesn't mean that the user's presumption should change.
What is at issue is the archiving of USENET messages, and archiving them isn't difficult. If you can't figure it out, you can even buy commercial off-the-self solutions to deal with it.
This is actually not much different from what was happening with BSD and AT&T UNIX: BSD needed to be cleaned before getting released.
USENET postings have expiration dates, which state to recipients how long the authors intends for them to be retained on public servers. Keeping postings any longer than that looks like a pretty clear violation of copyright. The argument that "you know when you post that..." is bogus. Music publishers also "know when they publish CDs" that their music will get copied, but that doesn't invalidate their copyright.
The fact is that DejaNews got away with this because they were big, and, hey, what could a random USENET poster do?
That's nonsense. There is no "work" involved in archiving USENET postings. DejaNews didn't manually classify or edit articles. What you need for USENET archiving is storage and a simple script. The whole USENET system is designed for easy, reliable replication and archiving. Yes, the storage costs money and it is "work" to buy new CDs and disk drives. But a "software pirate" doesn't acquire a copyright to the stuff he is copying just because it is takes time for him to copy the stuff.
Google should become more compliant with copyright law, not less. But, then, the company blatantly copies and retains other content as well ("cached pages").
It should be pretty obvious to the person posting it whether they have the copyright, and if they post anyway, well, as I was saying... that's the ethical side.
The reasoning is that free software and open source software is about principles, and it's bad to compromise those principles for some possible short-term gain.
From a purely practical point of view, I think it would just be a bad precedent for people to write such specific restrictions into licenses. Enforcement is difficult, and you can't easily revoke the restrictions when the company changes (many companies have come around to open source).
In this particular case, I don't even see anything particularly serious to get upset about. The guy is going after people who post commercial software to newsgroups. That would be a worthwhile thing to do even from the point of view of keeping some junk off of USENET.
From a technical point of view, someone who wants to post potentially incriminating material and uses NNTP from their cable modem deserves what they get.
As for the transparent proxies detecting content, unless there is legislation restricting users to a handfull of well-known formats, in the presence of even simple cryptography or compression, that's a pointless exercise. OTOH, I see no problem with companies trying to detect copyrighted works by participating in P2P networks; of course, since P2P is moving towards tit-for-tat schemes, these participants better be prepared to put up some interesting content themselves in order to particpate :-)
Aren't you glad to know that law.com has your privacy and consumer choice at heart? As long as you allow them to track you and use the latest version of Internet Explorer, you are fine.
While I'm not saying it will go that far, the logical end point of this might be that you go to a web site, type in the bits and pieces of stuff you want, and have your raw disk image updated regularly. User-visible client-side package management would become irrelvant, you'd be completely dependent on the subscription, and you could forget about CD-ROMs or other disconnected installations.
Yes, some form of convenient download and dependency checking is needed. And both RedHat and Ximian probably have the best intentions. But I'm not convinced that this is the right overall route to go in the long run.
In my experience, none of the automatic upgrades are entirely reliable. Debian is so dependent on automatic upgrades that packaging bugs get fixed more frequently, but problems occur even with Debian and can be very annoying. I see little reason to upgrade continuously. If it isn't broken, don't fix it.
As for the GPL license, I don't particularly like the policy behind it, and I think RMS's recent approach to using it is too heavy handed. But, then, I don't particularly like Microsoft's "we own your firstborn son" licenses either. The legal foundations on which they are based are similar, and if Allchin complains about the GPL, well, he might lead the way and get rid of Microsoft's heavy-handed licenses. See, what Allchin is really saying is that he is concerned when other people use copyright and software licensing laws in ways that interferes with Microsoft's business.
And he probably senses that economically open source software is a threat to his company because open source makes sense: it is finally a good mechanism to distribute software development costs equally, whereas all Microsoft can offer is having customers pay hundreds of dollars year after year for "improvements" they don't want in the first place.
Well, tough cookies, Allchin. Basic economic theory tells us that in an efficient market, Microsoft won't be making much in the way of profit. So, learn to live with it, and innovate a bit on the software side for a change.
The problem with tabular data is that it doesn't let you represent a lot of information in a convenient form: people often do need and want hierarchical/tree-structured data. And tabular data in UNIX isn't self-describing. Configuration files, package descriptions, bibliographic citatinos, etc. all need fairly complex descriptions.
There are many ways of representing tree structured data. XML wouldn't be my favorite, but it is workable. And XML is getting a fairly complete set of tools for dealing with tree structured data: search, extraction, restructuring, etc.
With this, maybe the Linux community will pick up some of that old UNIX spirit again. Today, the habit seems to be that when anything needs to get done on Linux, someone writes a big Gnome or command line program in C, or, on a good day, writes a monolithic Perl program. For example, something like "rpmfind" should really just be a collection of a few command line tools: something like the Xerces tools for extracting information, gunzip to uncompress the data, and curl to retrieve information. The same is true for a lot of other Linux applications.
Oh, if you want to play around with XML, I found some of the Apache Xerces tools at http://xml.apache.org/ to be quite useful. They come in both Java and C++ flavors.
The Sharp machines have always been a good value and a nice form factor. I think what has held them back was their hotsyncing software and limitations on programmability.
I think Mozilla, Gtk+, Qt, and the C++ libraries are so huge because people these days write for reuse and generality, not simplicity, maintainability, space, or efficiency. I have concluded it's pointless to argue that point. Most programmers educated in the last 10 years don't even believe that they are doing anything wrong when they do that.
Maybe games based on good narrative can be developed eventually. But until someone figures out how, game designers should perhaps just focus on making good games. There is ample precedent for how to go about doing that. And it neither involves making characters more "life like" nor adding a lot of bogus story background to the game.
Probably, you'll get the worst of both worlds: you buy the PC with a one-year subscription, to be activated when you first plug in the machine. You won't be able to get the PCs without those pre-paid subscriptions.
Besides, the PC and Windows platforms are so messy that things have to be preinstalled; the "put in CD and turn on" isn't just undesirable from a business point of view for Microsoft and PC vendors, it is technically hard.
If hardware vendors went back to selling plain hardware without any software, that would be great for Linux and Windows-alternatives. It's not going to happen.
The reason why many people believe that genes correspond to morphologies is because, despite the fact that genes code for proteins, it is empirically justified in a certain sense: genetic alterations in specific genes have been observed frequently to correspond to specific morphologies. Of course, that can be explained because those are the only alterations that are easily observed. It is likely that many other mutations have such widespread effects that they just cause death during development. And molecular biology has given us ample examples for the fact that many genes have many functions throughout development (e.g., "combinatorial codes").
All this has been abundantly clear to many molecular biologists for decades, including Gould's colleagues at Harvard. Unfortunately, there seems to be a class of scientists that make a name for themselves by stating that everybody else in their field is plain dumb and then restating the obvious as if they were the first to think of it. Sorry, but as a far as I'm concerned, while Gould expresses things very well, he hasn't said anything of interest in many years.
We may laugh this off, but this kind of nonsense has a corrosive effect. Even if I tell you that I'm going to tell you a lie, a false statement I make to you has been shown to influence your thinking and judgement later. Multiply that effect by several hundred million and who knows what happens? How much funding is NASA going to lose because of the false impressions this kind of show creates? The only thing I can see doing about this sort of thing is to express your outrage. If you subscribe to cable, unsubscribe and let them know why. When it comes to the dangers of media, this is the kind of stuff politicians should worry about.