Book review bottom line doesn't add up
on
jQuery Cookbook
·
· Score: 1
I didn't know what to make of this book review. The book review points out a lot of ways in which this book is a mess (poorly copy-edited, unclear motion from one author to another, terminology used without definition, etc.). Then at the end, it says:
Usable for both learning and reference, jQuery Cookbook delivers a tasty buffet of programming essentials, best practices, illustrative examples, optimization tips, and other information of value to JavaScript developers who wish to spice up their Web creations with jQuery.
What gives? This seems like a conclusion ripped from an entirely different review. If it's that this book is really the only choice right now, then you could say that. but this sentence here looks like it comes at the end of a rave review.
I confess to being a bit rusty on this subject, but isn't it the rationals that are a field? Or at the very least, you don't need to go to the reals to get a field, although they are a field...
The ISS does serve one useful purpose --- it gives support to the Russian space program. This, in my opinion, is a very good thing. I think we're all going to be a lot happier if we pay the people involved in that program to support the ISS, even if it's totally useless, instead of having them go off and work for North Korea, Iran, or whatever other power wants to acquire ballistic missiles, etc.
I just wish that the cost would be charged to disarmament, where it properly belongs, instead of sucking all the life out of the space program.
It sure seems to me that having a one-way ticket (or switching airlines in an itinerary) is a good way to bump your probability of being hassled. Does the TSA think that Islamic fundamentalists have a taboo against leaving a return trip ticket unused after a suicide bombing, so they are compelled to buy a more expensive one way ticket?;-)
Thanks. Good point. Static linking does go a long way to solving the problem with respect to libraries. But other problems remain, notably the invocation of infrastructure services that are separate programs, rather than libraries. E.g., we have a program that uses ImageMagick to generate images for display through a web interface. Since the ImageMagick command-line has not been stable, making our program work across different linux distros, which have made different decisions about what ImageMagick versions to package, made us miserable. It has just been easier to give our customers full linux distros, which we know and control, rather than trying to fuss with each customer's individual set-up (or give them a directory tree with the ancillary programs in known versions). The LiveCD approach also makes installation pretty much a non-issue.
But I think it says something that a LiveCD can be the simplest solution to distributing an application....
1. Is open source more complex? Answer: No, consider SAP.
2. Is it hard to develop for Linux deployment? Absolutely. There's too much worry about what libraries are going to be available on a particular distribution. At the company I work for, we have pretty much given up development for distribution to arbitrary linux systems. Now we put our software onto a LiveCD using Knoppix, so that we know the platform will have the right utilities, libraries, etc.
I'm a high-level language lover --- for work I mostly use Common Lisp, with some side trips into perl, prolog and domain specific languages that are ultra-high level. I do agree that these languages can let you get efficient use of programmer cycles.
But it's also worth keeping in mind Alan Perlis' old adage "Lisp programmers know the value of everything and the cost of nothing." All too often programmers, especially novice programmers, write terrible code in ultra high level languages, because they don't have good mental models of how the expressions in the languages are executed. Yes, the compiler can be clever, but for places where the compiler can't do the job, novices may get into trouble. For example, I have looked at Lisp code from a very respected source that kept a queue of objects that got to be 10,000s long, implemented as a list, and doing a lot of appends. This involves a simply horrific amount of pointer-chasing.
Programming environments that simulate non-determinism (e.g., prolog, many AI programs) can be even worse, since it's hard to envision how programs that search and backtrack will work.
Sounds like this might fly too low to present a danger to other aircraft, but the danger to people under the flight path may be profound, probably more than the incremental danger to privacy (since the LAPD already snoop extensively from the air).
The article describes them crashing the SkySeer in a flight test.
How do you think you would feel if the LAPD were to drop 3 kg on your head from 300 meters? Pretty bad, I would think.
UAVs have terrible safety records, and I don't feel great about putting them in human-flight airspace, much less flying them over a densely-populated city. This seems very irresponsible.
The way export is defined in US regulations and laws is not about sale. It has to do with making objects and information available. E.g., multinational companies are required to provide some segmentation in their computer networks to avoid exposing export controlled, or ITARS restricted information from reaching their non-US employees. Whether or not one thinks that the US government is becoming paranoid and over-secretive (I do), this is not an unreasonable definition of export. E.g., if one just gives centrifuges for enriching uranium to Iran, they are exported there, independent of whether one receives reimbursement. The arrival at the endpoint of the object or information is what the US government cares about, not whether someone is paid to supply the stuff. If you believe at all in export control, then it's not unreasonable for the US government to require that a vendor make some attempt to verify that its transactions comply with export control. Otherwise, you can just have someone say "I'd like to buy a whole lot of weapons-grade uranium. Here's my check."
I have had success with Mandriva 2006 on a Dell 700m. Wireless works, suspend to disk works (using Mandriva's multimedia kernel). It's a great Linux laptop.
The only thing that doesn't work on this machine is...Windows. Blue screens all the time on boot with bad device driver errors. Dunno if this is because there's some hardware error that Linux doesn't see, or the dual-boot causes issues.
Wish I could do without the Windows install, but I can't.
I use linux a lot, and I'm pretty comfortable. But I suck at using Windows. People who are real Windows mavens can be impatient and intolerant, too. Maybe there are just more Windows experts so the jerks are easier to sidestep and ignore.
I'm not convinced that there's some special form of grumpiness that comes with one OS or another...
For me, your "con" would be a pro, too. At the level of overcrowding and the way drivers seem always to be skating on the thin edge of road rage, I'm at the point of saying "bring on the robot drivers!" It's just not fun to drive any more. I despair of trains ever getting penetration in this country, and even if they do, who wants to sit next to a pack of knuckleheads yelling into their cell phones to show how important they are?
Robot cars? Intrusive? Where do I sign?
I want to set the destination on my dashboard, grab a newspaper, and kick back while my car does the work.
I don't think we're there yet in terms of the control systems, but it can't come too soon for me.
XUL is very good RAD tool. Much.. much much better than HTML.
For me, in order for something to qualify as a tool for rapid development, it must have (at least) the following:
An interpreter, allowing you to experiment with code snippets by typing them and seeing their effect immediately.
A debugger. The importance of this can be minimized by the presence of an interpreter, if the interpreter can access important parts of internal state.
Does anyone know how these XUL tools stack up along these dimensions?
I bought a copy of MacFarlane's Rapid Application Development Using Mozilla when it came out, and had a whack, but gave up because it was too hard to grope my way to successful development with what was available at that time. I had a brief correspondence with the late Mr. MacFarlane, asking him essentially these questions, and his response was, roughly speaking, "No, there isn't good support for experimental programming, but it's coming." Has it?
To make the question more concrete: As a point of comparison, how does the ease of incremental development with XUL these days compare with your favorite of perl + Tk or Gtk, Tcl/Tk, python + Tk or WxWidgets, or Visual Basic?
As long as you don't want to call up Microsoft for support, OEM software is just fine. But considering support rates ($35 a pop, or $245 for a professional incident), retail software may be a deal for those who lack basic troubleshooting skills, internet search capabilities, or impressionable tech-savvy relatives.
I've actually called Microsoft for support recently, when the activation process on a retail copy of XP Pro went south and the install destroyed itself. By the time the whole experience was over, I pretty near wanted to kill myself, and I was out a staggering number of hours. It never worked.
Free support like this doesn't sound like a deal to me... I would have been better off with no support, and not throwing more good time after bad.
It's like the old W.C. Fields joke: "First prize is a one-week vacation in Philadelphia." "Second prize is a two week vacation in Philadelphia."
Second prize is more paid-for Microsoft tech support.
This review is interesting, but very hard for the non-.NET indoctrinated to follow. Specifically, many terms go undefined, which is a particular problem for the review of a book aimed at MONO beginners. For example, what are: CLR IR GAC (I know what it stands for, but what does it do?) Since Chapter 7 is apparently about ADO.NET, it would be nice to give the reader a little sense of what this is.
On the contrary, this is not just old tactics on a new medium.
The international nature of this medium means that the Pentagon cannot prevent "blowback" from its propaganda into domestic US news media.
Compare this to the leaflets dropped from an airplane mentioned in another comment. Those leaflets were not going to come back from Nazi Germany to the United States, and even if they did, they were not going to be confused with news items written by journalists.
The Voice of America was explicitly intended not to be consumed by domestic US listeners.
The next question is "why do we care?" which seems to have been bruited about by a number of commenters.
The reason we (US citizens --- citizens of other nations will have different issues) should care is that as citizens of a democracy we have decision making power and responsibility. Further polluting our information environment with non-factual material is not going to improve the quality of our national decision-making, which is arguably already pretty badly contaminated by talk journalism, shock journalism, etc., etc.
IMO our national discourse about policy is already very badly polluted by wishful thinking masquerading as information-gathering. Adding Pentagon-generated materials for consumption in the Islamic world, for example, is not going to improve matters.
Note that I'm not issuing a blanket objection to propaganda. I'm just arguing about the dangers to rational thought that come from lying to oneself.
I think part of the problem here is that we still don't have a protocol for micropayments. Firms who spend money putting useful content on the web really only have three models for getting income, that I can see: 1. They take advertising. Attacking this revenue stream is what Nielsen is complaining about, IIUC. 2. They sell you a subscription. IMHO, this is never going to work. There's just too much to subscribe to, and each subscription costs too much. E.g., I get a lot of my interesting essay pointers through Arts and Letters Daily. They may point me to an article in Harper's. I am simply not going to buy a full year of a dead tree magazine, or even a subscription to their website, just to read an essay that interests me mildly. 3. They make you pay an arm and a leg for a single article. I pretty much never do this (technical articles might be an exception). If it's worth $5.00 + an annoying credit card interaction, I'd rather go to the public library and get it that way. AFAICT, in order for content suppliers like this to make money, there must be a protocol that makes it economically rewarding to collect small payments, and making a small payment must involve an absolute minimum number of keystrokes and web interactions. I don't really see how we can solve this problem with government regulation, unless we have some kind of tax on the web that is used to pay content providers. (I have a vague memory that some countries have a protocol like this for libraries, but don't really know.)
Other languages
on
Why Use GTK+?
·
· Score: 3, Interesting
One issue that I like to look at when I'm thinking about GUI toolkits is the question of what programming languages does it support.
I often need to rapidly prototype a graphical UI, and one that's not just a standard set of static attributes. I find that for these cases the graphical layout tools fall down pretty quickly, and I'm back to writing code to make the UI.
Now, if I'm going to be writing code for GUI prototypes, I want code that I can write, test, and show off fast. I don't want to start a language war, but to me that says "not C++."
So a big question for me is "what other languages does your toolkit support easily?" Is there a good perl interface? python? scheme? What can I use to lash it together quickly?
For this kind of thing, sadly enough, it seems like the venerable Tcl + Tk combination is still hard to beat.
And when you need an alternative language APIs, we need documentation that is native to those languages. All too many of these toolkits provide some rudimentary alternative UI, but it's just an export of the C++ API, and the programmer is expected to read the C++ documentation, and mentally convert that to the appropriate perl, python or what-have-you alternative.
So what are the easy cross-platform, scripting UI alternatives? Tcl + Tk, python + wxWidgets, and what else? Any way to get at those Swing libraries without heavy lifting with Java?
Given that a big theme of the article is that graphical IDEs are better than the debugger for almost all purposes, it would be nice to have a list of some of these perl-capable IDEs so that we can assess the claim.
I think it's of general interest when a major regional university like Tulane thinks it's a good idea to kill its engineering programs. This is another warning sign about the status of science and engineering in the United States. It's like a canary in a coal mine.
Science and engineering enrollment at the undergraduate level has been pretty feeble for a long time now.
I'd like to second this comment --- if you get your algorithm out of a book, RFC, or whatever, PLEASE GIVE A CITATION!
I recall trying to read a blob of code that was supposed to solve linear inequalities (simplex). But there are about a zillion ways to formulate this problem, how slack variables are used, etc. In the end, it was easier to rewrite it than reconstruct how it should be called.
Actually, this is not a perfect example. The author of that code had cited the book from which the algorithm had been taken. But it was published in the 1950s, was long out of print, and I couldn't even get it through inter library loan.
So keep that in mind if you think a citation is enough of a comment! If you got the algorithm from a freely-available (not copyrighted) source, consider committing the RFC or what-have-you into the source code repository (you do use one, right?) with the code.
This seems like a very odd quote to use to introduce this book, because one of the things Ranum seems to be talking about is that use of encryption and complex protocols can make security worse.
Why worse? Because the firewall, mainstay of our security efforts, becomes less and less effective. In the old days, your firewall could give a fairly cursory glance at packet headers, and have a good shot at catching the bad ones. Now, the packet header isn't so useful, because there is complex stuff inside the packet --- protocols are layered three deep or more.
That's why we need security at the application layer, instead of at the network transport layer --- the network transport layer just doesn't "know" enough to catch threats. What makes this really scary is that there is less of a bottleneck for the threats. It's nice as a defender to have a bottleneck you can protect. If the bottleneck goes away, and you have to protect all the applications, that's pretty scary.
Cryptography isn't going to give us a lot of help here, IMO. Yes, when our security has been breached, it can give a second line of defense, but that's about all (and even that seems a little suspect in a world with keystroke loggers).
It's somewhat ironic that a medium designed to share information has become one infested with what are essentially data and demographic mining sites like MySpace.
This should be a wake-up call, but probably won't be. I don't know what it will take to make people realize that a privacy policy that allows for arbitrary revision by the web site, and that allows for revision when the owning company is sold, is completely worthless.
Would you loan money to a company, with a contract that says "we'll pay, but we don't promise that any company that buys us will pay"?
It's nutso that we will do business with people that have this kind of policies, and it's nutso that these policies are allowed. Write your Congressional representatives and complain!
Slashdot Mirror
What gives? This seems like a conclusion ripped from an entirely different review. If it's that this book is really the only choice right now, then you could say that. but this sentence here looks like it comes at the end of a rave review.
I confess to being a bit rusty on this subject, but isn't it the rationals that are a field? Or at the very least, you don't need to go to the reals to get a field, although they are a field...
The ISS does serve one useful purpose --- it gives support to the Russian space program. This, in my opinion, is a very good thing. I think we're all going to be a lot happier if we pay the people involved in that program to support the ISS, even if it's totally useless, instead of having them go off and work for North Korea, Iran, or whatever other power wants to acquire ballistic missiles, etc.
I just wish that the cost would be charged to disarmament, where it properly belongs, instead of sucking all the life out of the space program.
It sure seems to me that having a one-way ticket (or switching airlines in an itinerary) is a good way to bump your probability of being hassled. ;-)
Does the TSA think that Islamic fundamentalists have a taboo against leaving a return trip ticket unused after a suicide bombing, so they are compelled to buy a more expensive one way ticket?
Thanks. Good point. Static linking does go a long way to solving the problem with respect to libraries. But other problems remain, notably the invocation of infrastructure services that are separate programs, rather than libraries. E.g., we have a program that uses ImageMagick to generate images for display through a web interface. Since the ImageMagick command-line has not been stable, making our program work across different linux distros, which have made different decisions about what ImageMagick versions to package, made us miserable. It has just been easier to give our customers full linux distros, which we know and control, rather than trying to fuss with each customer's individual set-up (or give them a directory tree with the ancillary programs in known versions). The LiveCD approach also makes installation pretty much a non-issue.
But I think it says something that a LiveCD can be the simplest solution to distributing an application....
1. Is open source more complex? Answer: No, consider SAP.
2. Is it hard to develop for Linux deployment? Absolutely. There's too much worry about what libraries are going to be available on a particular distribution. At the company I work for, we have pretty much given up development for distribution to arbitrary linux systems. Now we put our software onto a LiveCD using Knoppix, so that we know the platform will have the right utilities, libraries, etc.
I'm a high-level language lover --- for work I mostly use Common Lisp, with some side trips into perl, prolog and domain specific languages that are ultra-high level. I do agree that these languages can let you get efficient use of programmer cycles.
But it's also worth keeping in mind Alan Perlis' old adage "Lisp programmers know the value of everything and the cost of nothing." All too often programmers, especially novice programmers, write terrible code in ultra high level languages, because they don't have good mental models of how the expressions in the languages are executed. Yes, the compiler can be clever, but for places where the compiler can't do the job, novices may get into trouble. For example, I have looked at Lisp code from a very respected source that kept a queue of objects that got to be 10,000s long, implemented as a list, and doing a lot of appends. This involves a simply horrific amount of pointer-chasing.
Programming environments that simulate non-determinism (e.g., prolog, many AI programs) can be even worse, since it's hard to envision how programs that search and backtrack will work.
So, "yes," but also "yes, but."
Sounds like this might fly too low to present a danger to other aircraft, but the danger to people under the flight path may be profound, probably more than the incremental danger to privacy (since the LAPD already snoop extensively from the air).
The article describes them crashing the SkySeer in a flight test.
How do you think you would feel if the LAPD were to drop 3 kg on your head from 300 meters? Pretty bad, I would think.
UAVs have terrible safety records, and I don't feel great about putting them in human-flight airspace, much less flying them over a densely-populated city. This seems very irresponsible.
The way export is defined in US regulations and laws is not about sale. It has to do with making objects and information available. E.g., multinational companies are required to provide some segmentation in their computer networks to avoid exposing export controlled, or ITARS restricted information from reaching their non-US employees.
Whether or not one thinks that the US government is becoming paranoid and over-secretive (I do), this is not an unreasonable definition of export. E.g., if one just gives centrifuges for enriching uranium to Iran, they are exported there, independent of whether one receives reimbursement. The arrival at the endpoint of the object or information is what the US government cares about, not whether someone is paid to supply the stuff.
If you believe at all in export control, then it's not unreasonable for the US government to require that a vendor make some attempt to verify that its transactions comply with export control. Otherwise, you can just have someone say "I'd like to buy a whole lot of weapons-grade uranium. Here's my check."
I have had success with Mandriva 2006 on a Dell 700m. Wireless works, suspend to disk works (using Mandriva's multimedia kernel). It's a great Linux laptop.
The only thing that doesn't work on this machine is...Windows. Blue screens all the time on boot with bad device driver errors. Dunno if this is because there's some hardware error that Linux doesn't see, or the dual-boot causes issues.
Wish I could do without the Windows install, but I can't.
I use linux a lot, and I'm pretty comfortable. But I suck at using Windows. People who are real Windows mavens can be impatient and intolerant, too. Maybe there are just more Windows experts so the jerks are easier to sidestep and ignore.
I'm not convinced that there's some special form of grumpiness that comes with one OS or another...
For me, your "con" would be a pro, too. At the level of overcrowding and the way drivers seem always to be skating on the thin edge of road rage, I'm at the point of saying "bring on the robot drivers!" It's just not fun to drive any more. I despair of trains ever getting penetration in this country, and even if they do, who wants to sit next to a pack of knuckleheads yelling into their cell phones to show how important they are?
Robot cars? Intrusive? Where do I sign?
I want to set the destination on my dashboard, grab a newspaper, and kick back while my car does the work.
I don't think we're there yet in terms of the control systems, but it can't come too soon for me.
For me, in order for something to qualify as a tool for rapid development, it must have (at least) the following:
Does anyone know how these XUL tools stack up along these dimensions?
I bought a copy of MacFarlane's Rapid Application Development Using Mozilla when it came out, and had a whack, but gave up because it was too hard to grope my way to successful development with what was available at that time. I had a brief correspondence with the late Mr. MacFarlane, asking him essentially these questions, and his response was, roughly speaking, "No, there isn't good support for experimental programming, but it's coming." Has it?
To make the question more concrete: As a point of comparison, how does the ease of incremental development with XUL these days compare with your favorite of perl + Tk or Gtk, Tcl/Tk, python + Tk or WxWidgets, or Visual Basic?
I've actually called Microsoft for support recently, when the activation process on a retail copy of XP Pro went south and the install destroyed itself. By the time the whole experience was over, I pretty near wanted to kill myself, and I was out a staggering number of hours. It never worked.
Free support like this doesn't sound like a deal to me... I would have been better off with no support, and not throwing more good time after bad.
It's like the old W.C. Fields joke: "First prize is a one-week vacation in Philadelphia." "Second prize is a two week vacation in Philadelphia."
Second prize is more paid-for Microsoft tech support.
This review is interesting, but very hard for the non-.NET indoctrinated to follow. Specifically, many terms go undefined, which is a particular problem for the review of a book aimed at MONO beginners. For example, what are:
CLR
IR
GAC (I know what it stands for, but what does it do?)
Since Chapter 7 is apparently about ADO.NET, it would be nice to give the reader a little sense of what this is.
On the contrary, this is not just old tactics on a new medium.
The international nature of this medium means that the Pentagon cannot prevent "blowback" from its propaganda into domestic US news media.
Compare this to the leaflets dropped from an airplane mentioned in another comment. Those leaflets were not going to come back from Nazi Germany to the United States, and even if they did, they were not going to be confused with news items written by journalists.
The Voice of America was explicitly intended not to be consumed by domestic US listeners.
The next question is "why do we care?" which seems to have been bruited about by a number of commenters.
The reason we (US citizens --- citizens of other nations will have different issues) should care is that as citizens of a democracy we have decision making power and responsibility. Further polluting our information environment with non-factual material is not going to improve the quality of our national decision-making, which is arguably already pretty badly contaminated by talk journalism, shock journalism, etc., etc.
IMO our national discourse about policy is already very badly polluted by wishful thinking masquerading as information-gathering. Adding Pentagon-generated materials for consumption in the Islamic world, for example, is not going to improve matters.
Note that I'm not issuing a blanket objection to propaganda. I'm just arguing about the dangers to rational thought that come from lying to oneself.
I think part of the problem here is that we still don't have a protocol for micropayments. Firms who spend money putting useful content on the web really only have three models for getting income, that I can see:
1. They take advertising. Attacking this revenue stream is what Nielsen is complaining about, IIUC.
2. They sell you a subscription. IMHO, this is never going to work. There's just too much to subscribe to, and each subscription costs too much. E.g., I get a lot of my interesting essay pointers through Arts and Letters Daily. They may point me to an article in Harper's. I am simply not going to buy a full year of a dead tree magazine, or even a subscription to their website, just to read an essay that interests me mildly.
3. They make you pay an arm and a leg for a single article. I pretty much never do this (technical articles might be an exception). If it's worth $5.00 + an annoying credit card interaction, I'd rather go to the public library and get it that way.
AFAICT, in order for content suppliers like this to make money, there must be a protocol that makes it economically rewarding to collect small payments, and making a small payment must involve an absolute minimum number of keystrokes and web interactions.
I don't really see how we can solve this problem with government regulation, unless we have some kind of tax on the web that is used to pay content providers. (I have a vague memory that some countries have a protocol like this for libraries, but don't really know.)
One issue that I like to look at when I'm thinking about GUI toolkits is the question of what programming languages does it support.
I often need to rapidly prototype a graphical UI, and one that's not just a standard set of static attributes. I find that for these cases the graphical layout tools fall down pretty quickly, and I'm back to writing code to make the UI.
Now, if I'm going to be writing code for GUI prototypes, I want code that I can write, test, and show off fast. I don't want to start a language war, but to me that says "not C++."
So a big question for me is "what other languages does your toolkit support easily?" Is there a good perl interface? python? scheme? What can I use to lash it together quickly?
For this kind of thing, sadly enough, it seems like the venerable Tcl + Tk combination is still hard to beat.
And when you need an alternative language APIs, we need documentation that is native to those languages. All too many of these toolkits provide some rudimentary alternative UI, but it's just an export of the C++ API, and the programmer is expected to read the C++ documentation, and mentally convert that to the appropriate perl, python or what-have-you alternative.
So what are the easy cross-platform, scripting UI alternatives? Tcl + Tk, python + wxWidgets, and what else? Any way to get at those Swing libraries without heavy lifting with Java?
Given that a big theme of the article is that graphical IDEs are better than the debugger for almost all purposes, it would be nice to have a list of some of these perl-capable IDEs so that we can assess the claim.
I think it's of general interest when a major regional university like Tulane thinks it's a good idea to kill its engineering programs. This is another warning sign about the status of science and engineering in the United States. It's like a canary in a coal mine.
Science and engineering enrollment at the undergraduate level has been pretty feeble for a long time now.
You have an oversupply of engineers in the US if you prefer to buy them cheap from India and China. :-(
What that does for the national interest, I dunno, but multinational corporations are cutting themselves free from that kind of concern, anyway.
I'd like to second this comment --- if you get your algorithm out of a book, RFC, or whatever, PLEASE GIVE A CITATION!
I recall trying to read a blob of code that was supposed to solve linear inequalities (simplex). But there are about a zillion ways to formulate this problem, how slack variables are used, etc. In the end, it was easier to rewrite it than reconstruct how it should be called.
Actually, this is not a perfect example. The author of that code had cited the book from which the algorithm had been taken. But it was published in the 1950s, was long out of print, and I couldn't even get it through inter library loan.
So keep that in mind if you think a citation is enough of a comment! If you got the algorithm from a freely-available (not copyrighted) source, consider committing the RFC or what-have-you into the source code repository (you do use one, right?) with the code.
This seems like a very odd quote to use to introduce this book, because one of the things Ranum seems to be talking about is that use of encryption and complex protocols can make security worse.
Why worse? Because the firewall, mainstay of our security efforts, becomes less and less effective. In the old days, your firewall could give a fairly cursory glance at packet headers, and have a good shot at catching the bad ones. Now, the packet header isn't so useful, because there is complex stuff inside the packet --- protocols are layered three deep or more.
That's why we need security at the application layer, instead of at the network transport layer --- the network transport layer just doesn't "know" enough to catch threats. What makes this really scary is that there is less of a bottleneck for the threats. It's nice as a defender to have a bottleneck you can protect. If the bottleneck goes away, and you have to protect all the applications, that's pretty scary.
Cryptography isn't going to give us a lot of help here, IMO. Yes, when our security has been breached, it can give a second line of defense, but that's about all (and even that seems a little suspect in a world with keystroke loggers).
This should be a wake-up call, but probably won't be. I don't know what it will take to make people realize that a privacy policy that allows for arbitrary revision by the web site, and that allows for revision when the owning company is sold, is completely worthless.
Would you loan money to a company, with a contract that says "we'll pay, but we don't promise that any company that buys us will pay"?
It's nutso that we will do business with people that have this kind of policies, and it's nutso that these policies are allowed. Write your Congressional representatives and complain!