Re:Firefox versus Mozilla
on
Firefox Hacks
·
· Score: 1
I was particularly thinking of the installation hassles in terms of retraining my password manager.:-(
I haven't been able to tell if I can just pull its database over, somehow. I suspect not...
Re:Firefox versus Mozilla
on
Firefox Hacks
·
· Score: 1
Thanks. The plugins are a big benefit (greasemonkey looks particularly appealing), but that's only half the question when we're trying to analyze the decision. What's the cost? What do we lose? Can we carry over any of our configuration (aside from bookmarks)? If so, how much, and how do we do it?
Firefox versus Mozilla
on
Firefox Hacks
·
· Score: 1
I realize that this is a little off topic, and apologize.
Can anyone explain, to a longtime Mozilla user, what's the benefit to changing over to Firefox from Mozilla? I know it's supposed to start up faster, but that's not a huge issue, as far as I'm concerned.
Are there other reasons to change over?
If one does change over, how painful is the process? Any way to transfer configuration information? I figure the bookmarks are easy, but what about the Password manager? Are there resources to guide us through the process of changing over?
I don't know what your definition of "consulting" is, but in my book, it's a profession. If he had said up at the top that he was "volunteering," that would be one thing. But this is a clear statement that he was employed by the Dean campaign.
That's miles away from the scandal, and it's just plain disingenuous of the WSJ to make it seem otherwise. I've always liked their paper (but not the editorial page); I'm appalled and depressed by the way they're using the news section to editorialize here. Boo!
Actually, this is trivial. Lisp doesn't expose pointers to the programmer, either, but has long had an EQ (object/pointer equality) operator to complement its value equality operator, EQUAL.
Probably the more relevant prior art is in Lisp, where there are both EQ (same place in memory) and EQUAL (value/deep equality). These have been around for a dog's age.
And I don't think that having this be for BASIC should mean that this prior art isn't relevant. Given this prior art, surely the IsNot operator would be obvious to any reasonable practitioner, and this application should be biffed.
Is there any kind of public comment period that we can use to argue against this patent?
As a relative geezer, I see some irony in Microsoft having a spokesman who derides a competitor's product without trying it. This reminds me of the time when IBM lost its dominance to Microsoft. One of the things that led to their slump was their entirely inward-directed view. For example, developers were forced to use IBM source code editors, even if there were better tools; developers were not allowed to have electronic mail access to the outside world, except in special cases; and developers were trained in idiosyncratic IBM products, methods, etc. I recall a friend of mine, who had left an outside firm, showing his IBM colleagues alternative tools that left them speechless (PC development tools were pretty primitive in those days!). I don't think that Microsoft is succumbing to this at the lower levels of the firm. But it won't be good for them if they allow this to happen even for the suits.
The answer probably has a lot to do with our very low population density. Walking a lot is good for those of us who live in cities that are laid out acceptably for pedestrians. But what do you expect people to do when the nearest supermarket is > 1 mile away? What about the nearest doctor? hardware store? etc.
The US is simply not populated in a way that will allow us to give up cars, which must be a leading consumer of power.
No, we're going to have to find alternative power sources. We simply don't have a choice.
OK, this is a little off-topic, but if you're a European who wants to get Kerry elected, I would suggest not sending unsolicited emails and telephone calls.
...to open a second front against companies like this, and just refuse to buy their products?
If Blizzard is a necessity of life for you, like food, I guess you lose. But it seems like a luxury to me, and isn't it time to just refuse to give a company your money if you don't like what they do with it?
Most of the patent litigation that gets reported on slashdot is usually the other way around, heck most of it is just the potential of a patent to be used in a bad way, but there are cases where the little guy that poured his heart and soul into something was able to prevent a bigger company from ripping him off.
This is tragically rare. The problem is that patent litigation is staggeringly expensive. If you have a chance, read the story of the fellow who patented intermittent wipers. It took at least ten years of his life, full-time, to recover from the auto giants. Few of us can (or wish to) drop our entire lives and become patent avengers instead.
Patents have turned into things that large companies put into portfolios to trade with each other. These interlocking patents simply repel outsiders, and promote oligopoly.
I thought we were doing just fine when software could only be copyrighted, not patented. I long to have those days back.
I don't think this claim is accurate. For example, the provisions in connection with the FISA warrants radically change the search warrant rules, they don't just streamline them.
The other issue here is a question about whether the law is being properly applied. The Patriot Act did not proclaim itself as a law for streamlining search warrant, wiretap, and property seizure laws in general. It proclaimed itself as an anti-terrorism law. If it's being used to just catch Joe Scumbag, whether or not you thinking catching J.S. is a good thing, it's not a proper use of the law.
If the American people want a law that's a general streamlining of evidentiary rules, then they should get such a law. But they shouldn't get such a law under false pretenses, masquerading as an anti-terrorism law.
Actually, I don't agree with this claim; whining on/. about this case may be very important. Here's why:
If we're going to write to our Representatives and Senators about problems with the Patriot Act, we need to have a good understanding of the issues.
I haven't been able to discern the details yet, but if it is true that the Patriot Act was used to bust this guy, then it's an important thing for us to know about. Why? Because it gives us a concrete example to cite when we write our Senators and Representatives to say that the Patriot Act is being abused for non-counterterrorism purposes.
Note that there are a load of "if's" in the above! All we have so far is one person's assertion that the Patriot Act was abused to bust him. We need to get some kind of corroboration before using this example in letters. Letters citing this case could blow up in our faces if it turns out there was no Patriot Act abuse.
First another prior art item. The ilisp mode for emacs editing of common lisp has had a very similar mechanism for a long time. One could gradually build up a set of modified function definitions as a change set and compile them together. This is actually a very similar mechanism, underneath.
Actually, this seems like a simple extension to any number of literate-programming style comment preprocessors. One could almost do this just with the facilities offered by pod. [with a little evil scripting, one probably could do it with pod...]
What all this prior art suggests to me is not that this patent should be thrown out on the grounds of the prior art, but on the grounds of obviousness. Seems like any competent practitioner, offered a tool like grep and some scripting infrastructure, could come up with a similar facility.
I don't think we're in as much disagreement as you do.
I write tons of perl. Why? Because the vast majority of my code is not mission-critical. I'm not about to start writing even Ada, much less SPARK because, as you point out, it's not worth the cost.
The next question is whether B&D languages are worth the cost for mission-critical applications. I'm inclined to think that they are --- perhaps.
The "perhaps" has to do with what we take out of a language to make it a B&D language. There are lots of things I think are not worth removing. But I can think of some things that are a slam-dunk. For example, C and C++ have way too many ways to give aliases to a memory location. I'm inclined also to think that while C's type coercion is mostly benign, when you apply that liberal stance to C++, which has a far richer type system, it's a mistake. Making people responsible for turning floats into integers and vice versa doesn't seem too bad.
For that matter, unit discipline seems like a B&D thing that's way overdue. Losing the Mars probe is yet another wake-up call that we've slept through. Values should have units, and if you try to combine values with different units, the compiler should either barf, or add code to do the conversion.
The second "perhaps," is how do we evaluate these B&D languages? I'm actually not very optimistic about there being any quick answer from experimentation. There are way too many variables for us to expect quick answers from experimental studies any time soon. So we're going to have to muddle through on judgment. And in judgment, as you so clearly point out, we must weigh the costs against the benefits.
A last word about the costs, though. Right now we're only paying attention to the costs to the producer of the code. You point this out yourself when you say:
It's cheaper to write crappy software, unless your legal/political liability is so fscking high that you *have* to make in safe.
This is not a great way for a society to make decisions, and the plethora of spam, worms, and viruses makes that abundantly clear. Sooner or later, some high-stakes control application is going to sneak onto the internet, and then we'll really hurt. Another problem is that you and I may know that our code is not of a quality appropriate for a mission-critical application, but there's no reason to believe that the user shares this appreciation. There are all too many examples of people putting mission-critical applications on the same OS with applications that are not up to that standard. Then problems with the low-security applications (including the pests that plague them) get to have their chance at the mission-critical app. Think of the CIA director pulling classified files onto his Mac laptop and taking them home with him....
Somehow, we have to get a consideration for social costs into the decision-making of software producers. Maybe by fixing our liability system (software producers really don't have any appreciable legal liability now), maybe by regulation. I tend to favor the former, although that's very unfashionable these days. But we should be having that conversation now. When somebody blows up a nuclear power plant, refinery, or tanker full of LNG, it will be too late to have a rational discussion, and we'll get a bad answer.
Actually, I've seen a bunch of safety-critical systems that were written in Pascal or some godawful assembler, so I don't think complaining about B&D languages is a big issue for high-reliability systems.
That whole, "if this fails, people die," thing takes a lot of fun out of the process. You're just not going to add much back. And it's not at all clear you ought to. Perl is pretty far from a B&D language, but I'd sure hate to see an autopilot written in perl, no matter how productive or satisfied it made the coder.
The dominance of military research is nothing new in the U.S. The U.S., with its very strong belief in free market economics, has always had a hard time with federally-sponsored R&D. In the past, however, we've always done it, yet called the vast majority of it military, even though it often wasn't. Military research floated all boats, the way that space research does. [Similarly, we don't directly subsidize Boeing's production of airliners, the way the EU subsidizes Airbus, but we do give Boeing big contracts to build military aircraft.]
IMHO what has changed recently is that military research sponsors (notably DARPA) now call for very short-term turnaround in research results. Typically they like to see substantial results from a project in six months now. This means that there are new difficulties for using DARPA funding for basic research.
At the same time that military funding has been emphasizing short-term versus long term research, industrial research labs, and general industrial support for research, have collapsed. Essentially, corporate funders have been deterred by examples like Xerox PARC, Bell Labs, and IBM labs. They don't believe that corporate research generates results for the funding enterprise. This suggests that research must be funded as a social good, like highways, etc.
Unfortunately, military and enterprise funding for research has gone away at precisely a time when ideological sympathy for funding social goods through taxation is at an all-time low. And, of course, the federal budget is squeezed between tax cuts, recession, and the war effort. On the up-side, we don't have to balance the budget any more...:-)
Why does the Senate refuse to ratify the Land Mine Treaty? Because US and South Korean troops rely on Land Mines in case swarms of North Korean troops come pouring over the DMZ. In general, the US is much more likely to be facing a numerically superior enemy, but not a technically superior one. So the US military is likely to need land mines.
Why does the Senate refuse to ratify the Kyoto Accord? Beats the hell out of me, since all it was was a statement of warm and fuzzy intent, with no enforcement teeth or anything else that has to do with jobs. I figure that nobody felt like it was important enough to stick their necks out in the face of the Black Helicopter World Government paranoids. Anyway, I thought that was mostly the doing of the Prez, not the Senate.
Ummmm.... isn't it obvious that there's a place for code that will never be read? If I want to fix a bunch of my dang ID3 tags that are goofed, I'll just write some perl. And who cares if I don't understand it later? I could spend three weeks trying to figure out the zillion java classes to do it, writing comments in javadoc, etc.,etc., but that would be totally pointless. It's a scripting language, for goodness sake!
I'm curious --- why is it a big issue to have government funidng? Presumably the problem is the combination of government funding + proprietary? Off-hand, I would think that DARPA would like to see government-funded work used in the competition, because it would confirm that the work they were funding was good. On the other hand, there's always the danger that it would confirm that the work was no good, but they should be willing to face that too.
After all, the mission of DARPA is to support good research and development for the US government (and specifically the DoD). Its mission is not to sponsor fair competitions. Competitions may be a means to their end, but aren't the end itself. If their objectives are furthered by an unfair competition (as long as they aren't doing anything otherwise illegal or unethical), who cares?
Am I the only one who finds the new updater for XP really unhelpful?
Having been burned in the past, I configured the updater to just download the patches, but not install them, so that I can read the "details" before deciding whether to install the patch.
Clearly, Microsoft's definition of "details" diverges significantly from my own. Their detailed description always seems to be something like "There's a problem in application X that could allow an attacker to gain administrator privilege on your machine." Optionally, they might warn me that I won't be able to remove the patch once it's installed.
This is wildly insufficient. For one thing, if the patch is unremovable, the details should contain at least a capsule explanation of what the tradeoffs are likely to be --- in particular, whether or not installing this patch is likely to bust some beloved function. I still remember ruefully the time I installed a patch that busted synchronization of my WinCE handheld (I have since switched to a PalmOS device). I had to reinstall Windows to fix that one, and it cost me the better part of a work day.
The patch descriptions are also inadequate. E.g., the latest patch reports problem with FrontPage Server extensions. It's not even clear whether the problem is only if I'm running FrontPage server, or whether MS has just given a back door into my machine to any server that uses FrontPage.
I know, one can go to the Knowledge Base to get more details, but what part of "details" doesn't Microsoft understand? When I click on "details" I want details, not an opportunity to go yet further for the real details....
No, I'm sorry, I don't think we do want the Technomanifestos. This thread started as being about CS papers, not as being about rants. The papers here are mostly rants. The only scientific papers as such are the ones from before the 1960s (with the possible exception of Larry Wall's perl stuff). Most of these aren't even about computer science, per se (e.g., Eric Raymond's paper/book). Not to say these are bad. Just off-topic.
Slashdot Mirror
I was particularly thinking of the installation hassles in terms of retraining my password manager. :-(
I haven't been able to tell if I can just pull its database over, somehow. I suspect not...
Thanks. The plugins are a big benefit (greasemonkey looks particularly appealing), but that's only half the question when we're trying to analyze the decision. What's the cost? What do we lose? Can we carry over any of our configuration (aside from bookmarks)? If so, how much, and how do we do it?
I realize that this is a little off topic, and apologize.
Can anyone explain, to a longtime Mozilla user, what's the benefit to changing over to Firefox from Mozilla? I know it's supposed to start up faster, but that's not a huge issue, as far as I'm concerned.
Are there other reasons to change over?
If one does change over, how painful is the process? Any way to transfer configuration information? I figure the bookmarks are easy, but what about the Password manager? Are there resources to guide us through the process of changing over?
Thanks!
I don't know what your definition of "consulting" is, but in my book, it's a profession. If he had said up at the top that he was "volunteering," that would be one thing. But this is a clear statement that he was employed by the Dean campaign.
That's miles away from the scandal, and it's just plain disingenuous of the WSJ to make it seem otherwise. I've always liked their paper (but not the editorial page); I'm appalled and depressed by the way they're using the news section to editorialize here. Boo!
What's the big deal here? So it can be difficult dealing with vendors on Craigslist. Think it's easy to buy stuff through newspaper classifieds?
Actually, this is trivial. Lisp doesn't expose pointers to the programmer, either, but has long had an EQ (object/pointer equality) operator to complement its value equality operator, EQUAL.
Probably the more relevant prior art is in Lisp, where there are both EQ (same place in memory) and EQUAL (value/deep equality). These have been around for a dog's age.
And I don't think that having this be for BASIC should mean that this prior art isn't relevant. Given this prior art, surely the IsNot operator would be obvious to any reasonable practitioner, and this application should be biffed.
Is there any kind of public comment period that we can use to argue against this patent?
As a relative geezer, I see some irony in Microsoft having a spokesman who derides a competitor's product without trying it.
This reminds me of the time when IBM lost its dominance to Microsoft. One of the things that led to their slump was their entirely inward-directed view. For example, developers were forced to use IBM source code editors, even if there were better tools; developers were not allowed to have electronic mail access to the outside world, except in special cases; and developers were trained in idiosyncratic IBM products, methods, etc.
I recall a friend of mine, who had left an outside firm, showing his IBM colleagues alternative tools that left them speechless (PC development tools were pretty primitive in those days!).
I don't think that Microsoft is succumbing to this at the lower levels of the firm. But it won't be good for them if they allow this to happen even for the suits.
The answer probably has a lot to do with our very low population density. Walking a lot is good for those of us who live in cities that are laid out acceptably for pedestrians. But what do you expect people to do when the nearest supermarket is > 1 mile away? What about the nearest doctor? hardware store? etc.
The US is simply not populated in a way that will allow us to give up cars, which must be a leading consumer of power.
No, we're going to have to find alternative power sources. We simply don't have a choice.
OK, this is a little off-topic, but if you're a European who wants to get Kerry elected, I would suggest not sending unsolicited emails and telephone calls.
The UK Guardian tried that, and it doesn't seem to have been very successful. Have a look at http://www.guardian.co.uk/uselections2004/story/0, 13918,1332041,00.html.
...to open a second front against companies like this, and just refuse to buy their products?
If Blizzard is a necessity of life for you, like food, I guess you lose. But it seems like a luxury to me, and isn't it time to just refuse to give a company your money if you don't like what they do with it?
It's your money that's paying for their lawyers.
This is tragically rare. The problem is that patent litigation is staggeringly expensive. If you have a chance, read the story of the fellow who patented intermittent wipers. It took at least ten years of his life, full-time, to recover from the auto giants. Few of us can (or wish to) drop our entire lives and become patent avengers instead.
Patents have turned into things that large companies put into portfolios to trade with each other. These interlocking patents simply repel outsiders, and promote oligopoly.
I thought we were doing just fine when software could only be copyrighted, not patented. I long to have those days back.
I don't think this claim is accurate. For example, the provisions in connection with the FISA warrants radically change the search warrant rules, they don't just streamline them.
The other issue here is a question about whether the law is being properly applied. The Patriot Act did not proclaim itself as a law for streamlining search warrant, wiretap, and property seizure laws in general. It proclaimed itself as an anti-terrorism law. If it's being used to just catch Joe Scumbag, whether or not you thinking catching J.S. is a good thing, it's not a proper use of the law.
If the American people want a law that's a general streamlining of evidentiary rules, then they should get such a law. But they shouldn't get such a law under false pretenses, masquerading as an anti-terrorism law.
Actually, I don't agree with this claim; whining on /. about this case may be very important. Here's why:
If we're going to write to our Representatives and Senators about problems with the Patriot Act, we need to have a good understanding of the issues.
I haven't been able to discern the details yet, but if it is true that the Patriot Act was used to bust this guy, then it's an important thing for us to know about. Why? Because it gives us a concrete example to cite when we write our Senators and Representatives to say that the Patriot Act is being abused for non-counterterrorism purposes.
Note that there are a load of "if's" in the above! All we have so far is one person's assertion that the Patriot Act was abused to bust him. We need to get some kind of corroboration before using this example in letters. Letters citing this case could blow up in our faces if it turns out there was no Patriot Act abuse.
First another prior art item. The ilisp mode for emacs editing of common lisp has had a very similar mechanism for a long time. One could gradually build up a set of modified function definitions as a change set and compile them together. This is actually a very similar mechanism, underneath.
Actually, this seems like a simple extension to any number of literate-programming style comment preprocessors. One could almost do this just with the facilities offered by pod. [with a little evil scripting, one probably could do it with pod...]
What all this prior art suggests to me is not that this patent should be thrown out on the grounds of the prior art, but on the grounds of obviousness. Seems like any competent practitioner, offered a tool like grep and some scripting infrastructure, could come up with a similar facility.
I don't think we're in as much disagreement as you do.
I write tons of perl. Why? Because the vast majority of my code is not mission-critical. I'm not about to start writing even Ada, much less SPARK because, as you point out, it's not worth the cost.
The next question is whether B&D languages are worth the cost for mission-critical applications. I'm inclined to think that they are --- perhaps.
The "perhaps" has to do with what we take out of a language to make it a B&D language. There are lots of things I think are not worth removing. But I can think of some things that are a slam-dunk. For example, C and C++ have way too many ways to give aliases to a memory location. I'm inclined also to think that while C's type coercion is mostly benign, when you apply that liberal stance to C++, which has a far richer type system, it's a mistake. Making people responsible for turning floats into integers and vice versa doesn't seem too bad.
For that matter, unit discipline seems like a B&D thing that's way overdue. Losing the Mars probe is yet another wake-up call that we've slept through. Values should have units, and if you try to combine values with different units, the compiler should either barf, or add code to do the conversion.
The second "perhaps," is how do we evaluate these B&D languages? I'm actually not very optimistic about there being any quick answer from experimentation. There are way too many variables for us to expect quick answers from experimental studies any time soon. So we're going to have to muddle through on judgment. And in judgment, as you so clearly point out, we must weigh the costs against the benefits.
A last word about the costs, though. Right now we're only paying attention to the costs to the producer of the code. You point this out yourself when you say:
This is not a great way for a society to make decisions, and the plethora of spam, worms, and viruses makes that abundantly clear. Sooner or later, some high-stakes control application is going to sneak onto the internet, and then we'll really hurt. Another problem is that you and I may know that our code is not of a quality appropriate for a mission-critical application, but there's no reason to believe that the user shares this appreciation. There are all too many examples of people putting mission-critical applications on the same OS with applications that are not up to that standard. Then problems with the low-security applications (including the pests that plague them) get to have their chance at the mission-critical app. Think of the CIA director pulling classified files onto his Mac laptop and taking them home with him....
Somehow, we have to get a consideration for social costs into the decision-making of software producers. Maybe by fixing our liability system (software producers really don't have any appreciable legal liability now), maybe by regulation. I tend to favor the former, although that's very unfashionable these days. But we should be having that conversation now. When somebody blows up a nuclear power plant, refinery, or tanker full of LNG, it will be too late to have a rational discussion, and we'll get a bad answer.
Actually, I've seen a bunch of safety-critical systems that were written in Pascal or some godawful assembler, so I don't think complaining about B&D languages is a big issue for high-reliability systems.
That whole, "if this fails, people die," thing takes a lot of fun out of the process. You're just not going to add much back. And it's not at all clear you ought to. Perl is pretty far from a B&D language, but I'd sure hate to see an autopilot written in perl, no matter how productive or satisfied it made the coder.
The dominance of military research is nothing new in the U.S. The U.S., with its very strong belief in free market economics, has always had a hard time with federally-sponsored R&D. In the past, however, we've always done it, yet called the vast majority of it military, even though it often wasn't. Military research floated all boats, the way that space research does. [Similarly, we don't directly subsidize Boeing's production of airliners, the way the EU subsidizes Airbus, but we do give Boeing big contracts to build military aircraft.]
:-)
IMHO what has changed recently is that military research sponsors (notably DARPA) now call for very short-term turnaround in research results. Typically they like to see substantial results from a project in six months now. This means that there are new difficulties for using DARPA funding for basic research.
At the same time that military funding has been emphasizing short-term versus long term research, industrial research labs, and general industrial support for research, have collapsed. Essentially, corporate funders have been deterred by examples like Xerox PARC, Bell Labs, and IBM labs. They don't believe that corporate research generates results for the funding enterprise. This suggests that research must be funded as a social good, like highways, etc.
Unfortunately, military and enterprise funding for research has gone away at precisely a time when ideological sympathy for funding social goods through taxation is at an all-time low. And, of course, the federal budget is squeezed between tax cuts, recession, and the war effort. On the up-side, we don't have to balance the budget any more...
A couple of alternative interpretations:
Why does the Senate refuse to ratify the Land Mine Treaty? Because US and South Korean troops rely on Land Mines in case swarms of North Korean troops come pouring over the DMZ. In general, the US is much more likely to be facing a numerically superior enemy, but not a technically superior one. So the US military is likely to need land mines.
Why does the Senate refuse to ratify the Kyoto Accord? Beats the hell out of me, since all it was was a statement of warm and fuzzy intent, with no enforcement teeth or anything else that has to do with jobs. I figure that nobody felt like it was important enough to stick their necks out in the face of the Black Helicopter World Government paranoids. Anyway, I thought that was mostly the doing of the Prez, not the Senate.
Um, what is that supposed to mean?
Echoes of Jorge Luis Borges, who wrote in a story about a map of the world that was as big as the world itself....
Ummmm.... isn't it obvious that there's a place for code that will never be read? If I want to fix a bunch of my dang ID3 tags that are goofed, I'll just write some perl. And who cares if I don't understand it later? I could spend three weeks trying to figure out the zillion java classes to do it, writing comments in javadoc, etc.,etc., but that would be totally pointless. It's a scripting language, for goodness sake!
I'm curious --- why is it a big issue to have government funidng? Presumably the problem is the combination of government funding + proprietary? Off-hand, I would think that DARPA would like to see government-funded work used in the competition, because it would confirm that the work they were funding was good. On the other hand, there's always the danger that it would confirm that the work was no good, but they should be willing to face that too.
After all, the mission of DARPA is to support good research and development for the US government (and specifically the DoD). Its mission is not to sponsor fair competitions. Competitions may be a means to their end, but aren't the end itself. If their objectives are furthered by an unfair competition (as long as they aren't doing anything otherwise illegal or unethical), who cares?
Can you comment on this?
Am I the only one who finds the new updater for XP really unhelpful?
Having been burned in the past, I configured the updater to just download the patches, but not install them, so that I can read the "details" before deciding whether to install the patch.
Clearly, Microsoft's definition of "details" diverges significantly from my own. Their detailed description always seems to be something like "There's a problem in application X that could allow an attacker to gain administrator privilege on your machine." Optionally, they might warn me that I won't be able to remove the patch once it's installed.
This is wildly insufficient. For one thing, if the patch is unremovable, the details should contain at least a capsule explanation of what the tradeoffs are likely to be --- in particular, whether or not installing this patch is likely to bust some beloved function. I still remember ruefully the time I installed a patch that busted synchronization of my WinCE handheld (I have since switched to a PalmOS device). I had to reinstall Windows to fix that one, and it cost me the better part of a work day.
The patch descriptions are also inadequate. E.g., the latest patch reports problem with FrontPage Server extensions. It's not even clear whether the problem is only if I'm running FrontPage server, or whether MS has just given a back door into my machine to any server that uses FrontPage.
I know, one can go to the Knowledge Base to get more details, but what part of "details" doesn't Microsoft understand? When I click on "details" I want details, not an opportunity to go yet further for the real details....
No, I'm sorry, I don't think we do want the Technomanifestos. This thread started as being about CS papers, not as being about rants. The papers here are mostly rants. The only scientific papers as such are the ones from before the 1960s (with the possible exception of Larry Wall's perl stuff). Most of these aren't even about computer science, per se (e.g., Eric Raymond's paper/book). Not to say these are bad. Just off-topic.
Good pointer. But that's 13 articles! Any particularly good or bad?