One could envision a scenario in which Microsoft could embrace one of these libraries (e.g. from mono) and add their own code. They would be perfectly within their rights to do so.
Yes! Just look at what they did with Kerberos. Kerberos has an X11-type license, Microsoft extended it by adding functionality that would break machines running Kerberos code when interacting with machines running Microsoft-Kerberos. They documented the changes, but to get the documentation you had to agree to some ridiculous NDA which basically prevented you from revealing the changes, which would presumably include creating an open source version that uses those changes. This was a very big deal here on Slashdot several months ago.
Come on people - Microsoft has screwed up way too many times in the past to use the argument that "this time is different because they never screwed anybody over in this situation before." Chances are they have (as is the case here), and if they haven't, the screwing-over department is one of two departments where they innovate regularly (the other being their legal department). They have practically always had ulterior motives in the past when announcing things that sounded like they were turning over a new leaf (anybody remember how they said they wanted to fully support Java and free their developers when it first came out?) and it's unlikely that this time will be any different. I'm not touching Mono for as long as I can help it. (Sorry Miguel, it's not a comment on the quality of your software, I just fear what Microsoft has in store.)
You are operating from a faulty premise, which is that the record label must recover their production costs from sales.
...explanation of "loan sharking" snipped...
Don't the record companies make a killing on concerts as well? (Or is that just Ticketmaster?) The point being that concerts are yet another way for them to recover costs as well. It might actually account for a very substantial portion of their revenue (I'm just guessing here) and should also be factored in if you are calculating what money is available to the record companies to claim against production costs.
I think I speak for everyone when I say false positives are the only real hinderance to the filtering of spam.
This is one of the reasons I like SpamAssassin so much. Since it tags each message with a spam-ishness score you can set up tiered levels of filtering. After running Spamassassin for several weeks I ran some tests and discovered that none of my false positives ever scored above a 6, so I set up an extra filter for messages which score an 8 or higher that put them into an "almost definitely spam" folder. Everything from 5 - 8 goes into my "potential spam" folder and everything below 5 isn't tagged as spam at all. The bulk of my spam has actually been going into the "almost definitely spam" folder, which I only check every few days and any false positives I get have gone to my "potential spam" folder, which I check every few hours. My point is, it's pretty easy to set up SpamAssassin to work in accordance with to your tolerance for false positives and it can save you oodles of time even if your tolerance level is very low.
This guy is absolutely right and it's frightening to read all the responses along the lines of "nothing is perfect, so it makes no difference that Linux is being used." For those of you who think that nothing bad could happen, please read up on Therac 25. Perhaps the drill would "just shut off" if the power went off, but what if the user accidentally enters errorneous data which was not anticipated?
The Linux kernel was almost definitely not inteded for use in brain surgery. Frankly, if I were contributing to the kernel I would be very disturbed by this. Beforehand, if there was a bug in the kernel, somebody's webserver might crash every few months or maybe some credit card numbers could be stolen, but now bugs could potentially lead to death. I don't think I could live with that. Software that's used in life and death situations requires an entirely more rigorous method of design, implementation, and testing than goes on with the Linux kernel and I don't think it would be very comforting to the kernel contributors to know that they are now dealing with life and death.
Regardless of whether or not Linux would work here, the mere fact that they chose an OS that was not meant for a task like this would make me not trust the rest of their program. Apparently, there isn't a very wide understanding that things like this need to be designed differently. That's scary.
PS - I am in no way bashing Linux - I still recommend it for nearly all situtations that aren't life and death. Follow the link in my sig if you don't believe me.
What he says is at odds with the clauses of the LGPL, wherein it guarantees the right to reverse engineer the software linked to it--proprietary or no--and that any proprietary license that tries to revoke that right is in violation of the LGPL.
It sounds like you're talking about section 6 of the LGPL:
6. As an exception to the Sections above, you may also combine or
link a "work that uses the Library" with the Library to produce a
work containing portions of the Library, and distribute that work
under terms of your choice, provided that the terms permit
modification of the work for the customer's own use and reverse
engineering for debugging such modifications.
(I am not a lawyer, so the following is not legal advice.) I always thought that you could use LGPL'ed libraries with proprietary code as long as you allowed the LGPL'ed portions to be changed by the customer, and that the rest of the program could be restricted at will. I never really thought about your point that allowing reverse engineering of the program using the LGPL'ed library is a requirement as well, until you brought it up. I can see how you might get that from section 6, but I'm not sure that's entirely correct.
First of all, the LGPL only says that reverse engineering needs to be allowed to debug modification of the work for the customer's own use. It doesn't sound like reverse engineering for other purposes needs to be allowed. It also does not say that you need to allow any and all modifications, and in fact sections 6a and 6b specifically state what modifications need to be possible (essentially, being able to modify the LGPL'ed library). So I think this could be intrepretted to mean that you must allow the modifications stated in 6a or 6b and reverse engineering to debug those modifications, but that additional reverse engineering or modifications do not need to be permitted.
Furthermore, section 6 is an "exception" section anyway. My take on this is that section 6 does not apply if you are not distributing LGPL'ed code with your program (e.g., if you require the user to have a particular DLL or.so on his system beforehand), in which case I don't see any restriction on limitting reverse engineering.
Of course, I'm not a lawyer, so I could be off the mark on all this.
While
you're at it, point out an example of a site where it's useful.
You're looking at it. I use the "Top" link to navigate to the main page on Slashdot all the time. It comes in handy when I'm in the middle of a page and don't want to scroll around to find the "home" link on the page.
This sounds an awful lot like the Mozilla site navigation bar to me. This was removed right before 1.0 was released, which is why I'm still using a pre-1.0 version of Mozilla myself (well, that and the fact that Mozilla was already rock solid before 1.0).
Working on someone else's idea for nothing seems a particularly unproductive thing to do. Yes, you *might* get *some* future value (but probably not your fair share). You will almost certainly make yourself inelegible for unemployment benefits and you run the risk of getting caught up in the project without ever settling the question of proper remuneration.
I've always looked at equity this way too - why work for somebody else for 1, 2, or even 5 percent equity when you can work for yourself for 100% equity? Especially if you aren't getting paid, I can't envision any situation where it would make sense to accept anything less than 20% equity as the bulk of your compensation (certainly, it's OK to accept less equity if you are being paid a reasonbale rate, though). I would even say 20% is too low in most cases because for the same amount of risk you can have 100% of your own company.
Personally, my core competency is software design and development, so I do think there would be a benefit for me to join an existing company (for equity) where there are people with complementary skills, particularly in the business and sales areas. If I tried to do these other things myself, I probably wouldn't do as good of a job as others. The critical factor, however, is that even though I wouldn't do as good of a job at these other things, I would probably do a good enough job. The tradeoff is between having a small piece of a larger pie (when working with others who are better at selling, etc) or having the entirety of a smaller pie (when working alone). Say that a company were offering me 10% equity (which is unusually high from what I've seen) to work for them - do I really think that by working with others the product would be 900% more successful or that if I were to start my own company I would do 90% worse job at the non-technical aspects of things? The answer has always been "no" for me so far. Everybody should ask this question when considering equity as a substantial form of compensation and adjust the numbers accordingly.
Sorry guys, but this is seriously old news.. You can't call this a Christmass gift or anything. Our company's been using it for a year already (with a free license) and I personally have developed couple of plugins for it.
I think that the news that the author is referring to here is not about the release of Eclipse but that Eclipse can now be run using Gnu Classpath and gjc (which I doubt was the case a year ago). They are announcing this as a milestone for Classpath and gjc, not as a milestone for Eclipse. Eclipse just happens to be a hefty program that they are using as a test case, and its significance ends there. The significance of the announcement lies primarily with Classpath and gjc. That's my take on it anyway.
Very nice! Thanks. Instead of making all the Flash on the page totally disappear, is there anyway to use userContent.css to replace each flash occurrence with, say a big yellow rectangle, so I know that the page has been altered? I'm thinking that replacing the "none" in what you gave above with something else might do the trick, but I don't know what to put there. I'll gladly RTFM, just point me at the right FM (I didn't find a relevant one on Google).
actually, IBM has done something else, they have build SWT (standard window toolkit) and the footprint compaired to swing is by far better.
Thanks for the tip. I think I had looked at this awhile back, but at the time they only had support for Windows and Motif. My strategy for the peer to peer backup program I'm developing in Java has been to develop it in Swing first, since Swing is available on all platforms, and then go back and write native implementations for the platforms that need it. It's pretty easy to have multiple UIs if you separate things into model and view(s). Anyway, it sounds like the SWT could save me a bit of work on the second phase - thanks.
Java code cannot run with "superior" speed for the whole fucking Java is written in C (native) and therefore if it is faster, it is only because coders who coded Java were better than whoever coded the other stuff you are trying to compare to.
First of all, that's just not true. The JIT has access to runtime information about the code and it is therefore capable of making runtime decisions about how to best compile things. For instance, if you had a method called factorial(int) which was called frequenly with values entered by the user and the user happened to enter 20 a couple hundred times, the JIT can pick up on this and compile a special version of the method that returns a cached value for factorial(20). The HotSpot JVM supposedly does this, and I am not aware of any C compiler which does the same (it would be a much more difficult task, I would think). Yes, you could write your own caching code in C, but that is A) a lot more work and B) not likely to capture the common case as well as a JVM with runtime information can.
Secondly, as you hinted at, the design has a lot more to do with the speed of the code than what language it is written in. That's another factor in Java's favor when it comes to speed. If you know how to design things well, Java is one of the best languages to use because you spend most of the time implementing the design rather than the plumbing.
Nice. Java is a good concept. Slow, mind you, but good
It really isn't that slow anymore. I wrote an MD5 implementation in Java and when I tested the performance recently I was shocked that my Java implementation was significantly faster at checksumming a large file on my Linux system than the native md5sum binary. Now, Java wasn't entirely responsible for the speed increase (because I later rewrote parts as a native method and this was even faster still, so my optimizations had a lot to do with it), but my point is that JITs have gotten so good now that even computationally intensive Java code can run with comparable speed to native apps and sometimes with superior speed.
Now if only they would reduce the memory footprint of things like Swing.
Users hate scrolling left to right. Vertical scrolling seems to be okay, maybe because it's much more common.
Or how about because horizontal scrolling forces you to scroll once for every line you read while vertical scrolling only forces you to scroll once for every page you want to read? So, there's generally more than an order of mangitude less scrolling required with vertical scrolling than with horizontal scrolling.
I want someone to write a dummy flash plug in that does absolutely nothing, apart from stop the stupid 'do you want to download the latest Meglomedia Flash crud plug in...
I actually started on this one weekend and got to the point that you describe (my plug-in registered itself for the right mime types and then just didn't do anything when called). My original plan was to take it one step further and parse the Flash file on the web page, pull out all URLs, and display a list of URLs within the file so that I could easily get into those sites that annoyingly insist on having a Flash-only intro that require you to click on the intro to continue and which serve no purpose other than to look pretty. I haven't added this yet, though.
The reason that I stopped work on this when I did was because I ran into a dilema that I haven't thought of a good solution to. The problem is, there are many different types of sites that use flash:
Sites that use it only as a decoration but assume that you must have it (hence the the download prompt you describe).
Sites that use it only as a decoration and recognize this fact, gracefuly degrading when you view their page without flash.
Sites that "require" flash and won't let you in unless you have it regardless of how necessary it actually is.
Sites that have separate HTML and Flash versions where you are redirected to the appropriate section based upon whether you have Flash installed.
A plugin that masquerades as Flash would be great for all cases but the last one. I don't want to be automatically shuffled off to the Flash-only site when there is a nice HTML alternative that would work much better for me. Unfortunately, these sites are the ones where the webmasters use Flash responsibly, since they have alternatives available, so I don't really want to penalize them by breaking their site when it should work.
I think the solution might be to modify the nullplugin that comes with Mozilla. This is the default plugin, and I believe this is the piece of code responsible for the "do you want to download" messages. When it asks you if you want to download a plugin of a certain type, it should have a checkbox that says "don't ask me again" and then it should remember that mime type (come to think of it, I'd be happy if it never prompted me for any mime types, so maybe I should just disable the prompt globally). It would be nice if it also picked the URLs out of the file on the web page so that you could bypass annoying intros as well.
...Michael Wolff, a consultant who advised Time Warner not to buy AOL in the early 90's...
I'm pretty sure it was AOL that bought Time Warner (which is why it was such a shocker at the time). Not sure that it matters, though it does seem odd that Time Warner would be dictating to AOL considering they were the ones bought.
I've got $20 in my hand that I'd give to that effort in a second.
I'd pitch in $20 for a promising effort that wasn't ambiguously legal. How about if everybody pitched in $20 to pay for a top tier lawyer to start picking off (figuretively speaking) the top spammers one by one? The guy in the article was talking with glee about new spam software that sends you pop-ups when you aren't even using the web or email (I vaguely remember an article here on Slashdot about this being related to a hole in Windows Update or Windows Messenger or something like that on Windows). This sounds an awful lot like breaking into your computer - let's put the DMCA to good use for a change and sue these people into oblivion. How many people would pitch in for a legal fund to pursue this?
A safe deposit box, as somebody else mentioned, is a good idea in theory (and potentially in practice - I use one myself). However, the big problem that makes it and so many other backup strategies inadequate in practice is that it requires manual intervention and most people aren't going to use these mechanisms frequently enough. Most of the people I've talked with in small businesses backup maybe once a month (to tape, zip disk, CD, etc), if they're lucky. That's unfortunate because there's really no reason why you should be backing up any less frequently than every day your data changes...
I've been working on some software for awhile that provides a simple, automated solution. At the core it's peer to peer backup software, so all your computers will backup to each other. To protect against things like fire and theft, it also gives you the option to backup offsite to third party servers. So you could back up the bulk of your data via p2p (very fast, cheap, and easy) and backup your critical stuff offsite (still cheap and easy and potentially fast depending on the amount of data). I have a test version out now and I'm going to be releasing another test version in a few days, so email me if you're interested. And yes, there is a Linux version.
Java is not currently an option for the high-performance numerical and immersive graphical aspect of our applications.
Java isn't an all or nothing deal. You could write your app in Java and then convert the parts that really need performance into C and call it via JNI. Then you only have to deal with keeping a much smaller C library portable.
Just about every php.ini setting can be set right in your.htaccess file.
Hmmm... I did not know that. Thanks. That does make things a bit saner.
And no, magic_quotes_gpc affects exactly 0 functions.
Maybe not, but magic_quotes_runtime definitely does (I never said I was only talking about magic_quotes_gpc).
As for the class constructor thing. That's just a design decision. Just because you are used to one way of doing things doesn't mean a slightly different approach is wrong.
It also doesn't mean that the PHP approach is equally good - I would argue it's not. You weaken the abstraction barrier between the superclass and the subclass when you have to worry about whether or not to explicitly call the superclass' constructor. What happens when you decide that the superclass needs to be modified to do some initialization in its constructor when it didn't in the past? You need to go through and modify all subclasses to call the superclass constructor and you need to remember to call the constructor from any new subclasses you create. Conversely, implicitly calling the superclass constructor would automatically maintain the abstraction barrier without a significant downside, and thus would have been a better design decision.
I can't. All of my web sites are hosted through hosting services and I don't have control over the PHP config file. This is also a serious problem for people who want to write PHP libraries because they have to take into consideration whether all these little things with side effects like magic quotes and globals are on. What I've had to do is write my own functions which work the same regardless of what the config file says. This turns off magic quotes in essence, but it's very annoying that it was necessary.
Easy work around. Look up functions like is_numeric() or the ctype functions.
That's not an easy fix, you still have to retrofit your code everytime a bug pops up in order to get the benefits. In strongly typed languages the compiler will let you know right away that something is wrong letting you squash bugs before deployment, whereas with PHP you could have a bug that's dormant for a long period of time and when things do break it may not be immediately obvious where you need to add the checks you suggest.
"half baked OO support"
Nonsense. You obviously don't have a good understanding of how php objects work.
OK then, why don't constructors on PHP subclasses implicitly call the constructors of their superclasses?
Again nonsense. There are a lot of functions and that's a good thing but very few of them depend on some sort of a config setting.
Take a look at the settings in your/etc/php.ini file - there are a lot of them and quite a few of them alter the way that functions work. The magic quotes setting is far reaching and covers many functions. The stream wrapping settings are far reaching and cover many functions. Many of the other settings might effect one or two functions and in minor ways, but the subtelties can be your worst enemy at times because they are unexpected and sometimes unnoticed. And again, I cannot change my config file so I have to prepare for all contingencies as does anybody writing libraries.
Is there something inherently bad about PHP that should make me shy away from it, or is it more of a religious debate?
magic quotes (a truly awful idea)
register_globals (OK, this mistake will hopefully start to fade away now)
promiscuous casting and lack of type checking (debugging can be a nightmare as a result)
half baked OO support
Way too many functions do more than one thing or do different things depending on how your config file is set up. This is a pain to deal with unless you know for sure you are only going to be running your code on servers you set up and configure forever. If you are using a hosting service, it can making moving to a different service far more problematic than it needs to be.
Other people have said lack of separation between content and logic is a problem, but that's not a limitation of the language and is actually pretty easy to work around with a simple template library.
People have also said that there is no standardized DB support. That's not true. There is an ODBC module that you can use to make ODBC calls from PHP and this works fine (I've used it on Linux and BSD myself).
Despite all the reasons I listed above for why PHP irks me at times, I still use PHP for most web apps because:
It is very widely supported.
It runs on OpenBSD unlike any Java app servers (last I checked anyway, which was admittedly awhile ago)
It's open source
It's quite stable
The bottom line is that it's not perfect but it's still the best tool for a lot of jobs.
Slashdot Mirror
Come on people - Microsoft has screwed up way too many times in the past to use the argument that "this time is different because they never screwed anybody over in this situation before." Chances are they have (as is the case here), and if they haven't, the screwing-over department is one of two departments where they innovate regularly (the other being their legal department). They have practically always had ulterior motives in the past when announcing things that sounded like they were turning over a new leaf (anybody remember how they said they wanted to fully support Java and free their developers when it first came out?) and it's unlikely that this time will be any different. I'm not touching Mono for as long as I can help it. (Sorry Miguel, it's not a comment on the quality of your software, I just fear what Microsoft has in store.)
The Linux kernel was almost definitely not inteded for use in brain surgery. Frankly, if I were contributing to the kernel I would be very disturbed by this. Beforehand, if there was a bug in the kernel, somebody's webserver might crash every few months or maybe some credit card numbers could be stolen, but now bugs could potentially lead to death. I don't think I could live with that. Software that's used in life and death situations requires an entirely more rigorous method of design, implementation, and testing than goes on with the Linux kernel and I don't think it would be very comforting to the kernel contributors to know that they are now dealing with life and death.
Regardless of whether or not Linux would work here, the mere fact that they chose an OS that was not meant for a task like this would make me not trust the rest of their program. Apparently, there isn't a very wide understanding that things like this need to be designed differently. That's scary.
PS - I am in no way bashing Linux - I still recommend it for nearly all situtations that aren't life and death. Follow the link in my sig if you don't believe me.
First of all, the LGPL only says that reverse engineering needs to be allowed to debug modification of the work for the customer's own use. It doesn't sound like reverse engineering for other purposes needs to be allowed. It also does not say that you need to allow any and all modifications, and in fact sections 6a and 6b specifically state what modifications need to be possible (essentially, being able to modify the LGPL'ed library). So I think this could be intrepretted to mean that you must allow the modifications stated in 6a or 6b and reverse engineering to debug those modifications, but that additional reverse engineering or modifications do not need to be permitted.
Furthermore, section 6 is an "exception" section anyway. My take on this is that section 6 does not apply if you are not distributing LGPL'ed code with your program (e.g., if you require the user to have a particular DLL or .so on his system beforehand), in which case I don't see any restriction on limitting reverse engineering.
Of course, I'm not a lawyer, so I could be off the mark on all this.
This sounds an awful lot like the Mozilla site navigation bar to me. This was removed right before 1.0 was released, which is why I'm still using a pre-1.0 version of Mozilla myself (well, that and the fact that Mozilla was already rock solid before 1.0).
Personally, my core competency is software design and development, so I do think there would be a benefit for me to join an existing company (for equity) where there are people with complementary skills, particularly in the business and sales areas. If I tried to do these other things myself, I probably wouldn't do as good of a job as others. The critical factor, however, is that even though I wouldn't do as good of a job at these other things, I would probably do a good enough job. The tradeoff is between having a small piece of a larger pie (when working with others who are better at selling, etc) or having the entirety of a smaller pie (when working alone). Say that a company were offering me 10% equity (which is unusually high from what I've seen) to work for them - do I really think that by working with others the product would be 900% more successful or that if I were to start my own company I would do 90% worse job at the non-technical aspects of things? The answer has always been "no" for me so far. Everybody should ask this question when considering equity as a substantial form of compensation and adjust the numbers accordingly.
I wonder where they got that idea. The resemblance is striking.
Secondly, as you hinted at, the design has a lot more to do with the speed of the code than what language it is written in. That's another factor in Java's favor when it comes to speed. If you know how to design things well, Java is one of the best languages to use because you spend most of the time implementing the design rather than the plumbing.
Now if only they would reduce the memory footprint of things like Swing.
The reason that I stopped work on this when I did was because I ran into a dilema that I haven't thought of a good solution to. The problem is, there are many different types of sites that use flash:
- Sites that use it only as a decoration but assume that you must have it (hence the the download prompt you describe).
- Sites that use it only as a decoration and recognize this fact, gracefuly degrading when you view their page without flash.
- Sites that "require" flash and won't let you in unless you have it regardless of how necessary it actually is.
- Sites that have separate HTML and Flash versions where you are redirected to the appropriate section based upon whether you have Flash installed.
A plugin that masquerades as Flash would be great for all cases but the last one. I don't want to be automatically shuffled off to the Flash-only site when there is a nice HTML alternative that would work much better for me. Unfortunately, these sites are the ones where the webmasters use Flash responsibly, since they have alternatives available, so I don't really want to penalize them by breaking their site when it should work.I think the solution might be to modify the nullplugin that comes with Mozilla. This is the default plugin, and I believe this is the piece of code responsible for the "do you want to download" messages. When it asks you if you want to download a plugin of a certain type, it should have a checkbox that says "don't ask me again" and then it should remember that mime type (come to think of it, I'd be happy if it never prompted me for any mime types, so maybe I should just disable the prompt globally). It would be nice if it also picked the URLs out of the file on the web page so that you could bypass annoying intros as well.
A safe deposit box, as somebody else mentioned, is a good idea in theory (and potentially in practice - I use one myself). However, the big problem that makes it and so many other backup strategies inadequate in practice is that it requires manual intervention and most people aren't going to use these mechanisms frequently enough. Most of the people I've talked with in small businesses backup maybe once a month (to tape, zip disk, CD, etc), if they're lucky. That's unfortunate because there's really no reason why you should be backing up any less frequently than every day your data changes...
I've been working on some software for awhile that provides a simple, automated solution. At the core it's peer to peer backup software, so all your computers will backup to each other. To protect against things like fire and theft, it also gives you the option to backup offsite to third party servers. So you could back up the bulk of your data via p2p (very fast, cheap, and easy) and backup your critical stuff offsite (still cheap and easy and potentially fast depending on the amount of data). I have a test version out now and I'm going to be releasing another test version in a few days, so email me if you're interested. And yes, there is a Linux version.
Java isn't an all or nothing deal. You could write your app in Java and then convert the parts that really need performance into C and call it via JNI. Then you only have to deal with keeping a much smaller C library portable.
- magic quotes (a truly awful idea)
- register_globals (OK, this mistake will hopefully start to fade away now)
- promiscuous casting and lack of type checking (debugging can be a nightmare as a result)
- half baked OO support
- Way too many functions do more than one thing or do different things depending on how your config file is set up. This is a pain to deal with unless you know for sure you are only going to be running your code on servers you set up and configure forever. If you are using a hosting service, it can making moving to a different service far more problematic than it needs to be.
Other people have said lack of separation between content and logic is a problem, but that's not a limitation of the language and is actually pretty easy to work around with a simple template library.People have also said that there is no standardized DB support. That's not true. There is an ODBC module that you can use to make ODBC calls from PHP and this works fine (I've used it on Linux and BSD myself).
Despite all the reasons I listed above for why PHP irks me at times, I still use PHP for most web apps because:
- It is very widely supported.
- It runs on OpenBSD unlike any Java app servers (last I checked anyway, which was admittedly awhile ago)
- It's open source
- It's quite stable
The bottom line is that it's not perfect but it's still the best tool for a lot of jobs.