You are aware that Caml is implemented in C, a language that has most of C++'s problems with a lot of the safety bits removed.
This shows you have a very poor understanding of how compilers work. GCC's C++ compiler is written in C.
GCJ (Java compiler) is written in C. So what?
In fact you are factually wrong anyway - C is used to boot a mini-Caml compiler which is
then used to compile the main compiler, written mostly in Caml.
Firefox is actually written in javascript, isn't that a modern language? The libraries firefox uses are written in C++, [...]
Small parts of it are written in Javascript (the "chrome" - ie. the rules for automating actions in the user interface), but the vast majority is C++, including the crucial rendering engine which in Moz/FF renders web pages and chrome.
Would use of a functional language provide any tangible benefit to a web browser?
The code would be shorter (so easier to modify / verify) and wouldn't be subject to
buffer overflows or most types of memory leak. These seem fairly worthwhile. On the other hand fewer people are familiar
with functional languages (despite attempts by people like me to continually evangelise them!) so that might put off potential contributors.
The OCAML garbage collector isn't particularly suitable for a multithreaded application,
Very true. Firefox isn't multithreaded though, it's event-driven. I'm not sure why threads
are really needed for a web browser. Multiple processes might make some sense,
to provide real isolation between web sites / or web browser code and plug-ins, but it
would cost some speed.
I must pick you up on one thing. A smart pointer is a good thing - and about the best compromise you can get if you pass pointers around and can't reliably predict their lifetime (if you can predict it of course the standard alloc/free is far more efficient). A properly written one is *not* subject to threading issues... the one in the article merely sucked.
The central assumption of C/C++ is that once allocated, memory cannot be moved (because there are or could be unknown pointers to that memory anywhere). If you remove that assumption (as a high quality memory-compacting garbage collector does) then you'll see that actually smart pointers are really a very inefficient type of garbage collection, commonly known as reference counting. The problems with reference counting are well known.
Garbage collectors have the same problem that FF has - they eat memory.. they're basically big caches (if you think FF is bad try running a largish java app for any length of time - I've seen one take out a server with 2gb of ram!).
Just because Java is a crappy language with a poor implementation does not mean that all garbage collectors are bad.
Emacs also has a terrible garbage collector, and I guess because a lot of coders use emacs and have waited in front of it while it hangs doing a collection, they get the impression that all garbage collectors must therefore be bad. There are excellent GCs around for functional languages, which make them perform on a par (and sometimes faster) than tightly coded C programs.
Actually I'd like to see a browser written in a real modern language (I'm sure you were being sarcastic and are fully aware that Java is not a modern language). How about ocaml-fox?
5). Memory leak: I often run Firefoxes for a whole week long. Yes, you read that correctly. I often just leave important links open when I leave work, then I login back from home and continue useing it, then again tomorrow from work, and so on. After a week it often eats up around half a gigs of memory, true. But really, how many of you do such things ?
Sorry, I'm gonna rant now...
What, a whole week? My computer (running Ubuntu 6.06LTS) is up all the time. Basically it goes down when there's a power cut. Current uptime is 55 days (reflects the fact that I moved house 55 days ago). X hasn't been restarted in all that time.
Why does Firefox need so much memory to display a few web pages? And why doesn't it at the very least return the memory when I, say, close all but one web page? I mean, I know C++ is an awful computing language to write
anything in, and modern languages have garbage collectors, but still there's enough memory checking tools out there
so they can catch these memory leaks by now.
Please, by all means write a forum BBCode parser that outputs valid XHTML that works under the application/xml+html mime-type.
The wiki I wrote does something analogous to that. (Not BBcode, but MediaWiki-ish markup which is similar). It's not really so hard for a competent programmer.
I agree google analytics is helpful. However, it also increased my page load times by a long shot, so it had to be removed. It simply doesn't serve fast enough.
Did you try putting the Javascript somewhere other than the <head>? Obviously that's the recommended place, but in fact most of the functionality still works if you bury the Javascript down as close to </body> as it will go, and that should have less effect on the effective[1] page load time.
Rich.
[1] By "effective" I mean the time until the browser can render the page for the user, rather than the total load time.
Why was the parent marked "Off Topic"? The comment is absolutely right - this service has nothing whatsoever to do with HTML validation. It's a technique for getting more sales on your site by testing different combinations of headings / content in adverts to find out what works best.
Unless the web sites get indexed by Google, and zombies use specially chosen keywords to search for their latest encrypted instructions.
Could be slow. Better for the botnet herders to buy Google AdWords adverts on obscure keywords using a stolen credit card, or make use of Google's free $50 coupons when they offer them.
I'm sure I'm the only person on Earth who ever has to do this with any photo so I guess I can completely understand why "people" never what to do this.
You're not the only person on earth who wants to do this with the GIMP. For
work I take screenshots of web pages and have to annotate them (text, rectangles, arrows) to show people
what's wrong. I'd love to do this with the GIMP (I already use the GIMP for
cropping, etc.), but I have no idea.
I use GIMP quite a bit, although I'm by no means any sort of expert.
Please add the ability to do circles, lines, arrows, rectangles and so on in an obvious way.
Just the other day I had a screenshot which I was manipulating in GIMP, and all
I wanted to do was to add an arrow pointing to some specific feature (it was a screenshot
of a Google search, and I wanted to add an arrow pointing to the search box). Can
this be done? Probably. Can this be done by me who's been using GIMP since
the late 90s? No.
Its stupid comments like that that give OSS a bad name. Perhaps he's running
a business full time and hasn't got the time/ability to fix it himself or the
money to pay someone. Come down out of Geek Cuckoo Land and smell the real
world.
Yeah, and while we're at it perhaps Microsoft should just give
away all their software to people who have no money to pay for
it.... Right.
I didn't want to move to Firefox 1.5. It worked slower and ate far mre RAM on my machines.
But alas, few exploits later, I updated.
Because Firefox 1.0 support ended the moment 1.5 was out.
No it didn't. Official support for which you never paid a penny may have ended,
but there was nothing to stop you taking the code & fixing it yourself, or
paying a programmer to do it for you.
The conundrum is that [third party open source developers, eg. Debian, are] trustworthy... right up until they screw you over.
Well you don't need to trust them. If you've got the source
you can just look at the source and the patch (and even the vulnerability
if it was a full-disclosure list) and check it for yourself.
Or if you're not a competent programmer, pay a programmer on your
behalf to do the check.
In many ways its funny to see the Windows closed-source-is-best
Microsoft-is-always-right "community" acting this way. They
can't see that it's obviously better in this case to have source,
and third-party patches (with source) are clearly better for everyone.
My only defense is that the bubble hadn't even come to Texas yet, and nobody thought stock was worth the risk of working for a little startup.
Employees routinely get screwed even if they have good stock options. It's very common for the company - just before sale/IPO - to unilaterally change the terms of the deal so that employees are locked in for much longer than they agreed to be.
FWIW I once had options for 1% of a company which was later valued at (sticks pinky in mouth) 1 billllion dollars. I left before the IPO and before it was worth huge $$$, but the others I worked with got screwed in the above way and were only able to get out after the dot com boom had collapsed and the company was worth nothing.
Strangely enough the founders/directors didn't change their own terms and walked away with millions.
Who are these people who change databases all the time? We just figured that PostgreSQL was the best, it's free, we know how to use it and tune it, and so that's all we ever use. Now we can use all the wonderful SQL features of PostgreSQL and not worry.
A solution which forced us to regress to using Java or.Net is no solution at all BTW.
Well, it's hardly the first time that Sun has got involved in scripting/dynamic languages.
Back in 1994, Sun hired the core developer behind Tcl/Tk, and asked him to form a team around the language / graphical toolkit.
The toolkit was very widely used and quite promising (for the time), but
it languished at Sun and eventually they dumped it.
So while I firmly believe that copying software illegal for personal use is a minor civil infraction, like speeding, and should be punished accordingly (a small fine that's enough to make you not want to do it but proportional to the harm) I believe that commercial copyright infringement is much more serious.
Where to start? Speeding endangers not just yourself but other road users and pedestrians.
The amount of energy in your car increases with the square of speed making stopping
harder and your car more deadly when it hits someone. Rules about speeding are
one of the ways to prevent injury and death (rules against careless or dangerous
driving, and driving under the influence being others).
In the UK speeding is very much not a minor crime. People can be and
are routinely banned from driving and imprisoned for speeding.
The "war on copying" is something like the modern "war on drugs".
Computers are all about copying - it's their most basic operation.
Networks have brought the price of copying down to almost nothing.
The question becomes, in a world where distribution costs nothing,
how to make sure people keep creating stuff? Should that be
through the old industries who controlled distribution very
tightly? Well, probably not. In the arts, it turned out that
those industries were both very bad at producing works of art, and
they didn't compensate the artists very well either. In software
it seems that for whatever reason people will write the software
for free and live by selling support.
Now I'm no more going to defend this guy who was caught selling
millions of $$$s of software illegally. He's rather like a crack
dealer. I may think that coca-derived products should be
legalised (ie. regulated), but I won't defend someone who makes his living in
the current unregulated environment selling crack to anyone.
Slashdot Mirror
You are aware that Caml is implemented in C, a language that has most of C++'s problems with a lot of the safety bits removed.
This shows you have a very poor understanding of how compilers work. GCC's C++ compiler is written in C. GCJ (Java compiler) is written in C. So what? In fact you are factually wrong anyway - C is used to boot a mini-Caml compiler which is then used to compile the main compiler, written mostly in Caml.
Rich.
Firefox is actually written in javascript, isn't that a modern language? The libraries firefox uses are written in C++, [...]
Small parts of it are written in Javascript (the "chrome" - ie. the rules for automating actions in the user interface), but the vast majority is C++, including the crucial rendering engine which in Moz/FF renders web pages and chrome.
Would use of a functional language provide any tangible benefit to a web browser?
The code would be shorter (so easier to modify / verify) and wouldn't be subject to buffer overflows or most types of memory leak. These seem fairly worthwhile. On the other hand fewer people are familiar with functional languages (despite attempts by people like me to continually evangelise them!) so that might put off potential contributors.
The OCAML garbage collector isn't particularly suitable for a multithreaded application,
Very true. Firefox isn't multithreaded though, it's event-driven. I'm not sure why threads are really needed for a web browser. Multiple processes might make some sense, to provide real isolation between web sites / or web browser code and plug-ins, but it would cost some speed.
Rich.
I must pick you up on one thing. A smart pointer is a good thing - and about the best compromise you can get if you pass pointers around and can't reliably predict their lifetime (if you can predict it of course the standard alloc/free is far more efficient). A properly written one is *not* subject to threading issues... the one in the article merely sucked.
The central assumption of C/C++ is that once allocated, memory cannot be moved (because there are or could be unknown pointers to that memory anywhere). If you remove that assumption (as a high quality memory-compacting garbage collector does) then you'll see that actually smart pointers are really a very inefficient type of garbage collection, commonly known as reference counting. The problems with reference counting are well known.
Garbage collectors have the same problem that FF has - they eat memory.. they're basically big caches (if you think FF is bad try running a largish java app for any length of time - I've seen one take out a server with 2gb of ram!).
Just because Java is a crappy language with a poor implementation does not mean that all garbage collectors are bad. Emacs also has a terrible garbage collector, and I guess because a lot of coders use emacs and have waited in front of it while it hangs doing a collection, they get the impression that all garbage collectors must therefore be bad. There are excellent GCs around for functional languages, which make them perform on a par (and sometimes faster) than tightly coded C programs.
Rich.
Actually I'd like to see a browser written in a real modern language (I'm sure you were being sarcastic and are fully aware that Java is not a modern language). How about ocaml-fox?
Rich.
5). Memory leak: I often run Firefoxes for a whole week long. Yes, you read that correctly. I often just leave important links open when I leave work, then I login back from home and continue useing it, then again tomorrow from work, and so on. After a week it often eats up around half a gigs of memory, true. But really, how many of you do such things ?
Sorry, I'm gonna rant now ...
What, a whole week? My computer (running Ubuntu 6.06LTS) is up all the time. Basically it goes down when there's a power cut. Current uptime is 55 days (reflects the fact that I moved house 55 days ago). X hasn't been restarted in all that time.
Why does Firefox need so much memory to display a few web pages? And why doesn't it at the very least return the memory when I, say, close all but one web page? I mean, I know C++ is an awful computing language to write anything in, and modern languages have garbage collectors, but still there's enough memory checking tools out there so they can catch these memory leaks by now.
Or is it a misguided attempt to "cache" stuff in memory, which is about the stupidist thing you can do given that today memory is very slow versus processors, so usually it's faster just to recompute what you need when it's needed. Typical C++ programmers wouldn't know that though - they're still reimplementing reference counting on every one of their classes, when a central, optimized garbage collector would be a lot faster.
/rant over
Rich.
Please, by all means write a forum BBCode parser that outputs valid XHTML that works under the application/xml+html mime-type.
The wiki I wrote does something analogous to that. (Not BBcode, but MediaWiki-ish markup which is similar). It's not really so hard for a competent programmer.
Rich.
One of the advantages is that you could use XSLT to convert an XHTML document in something else (like LaTeX, RTF, ...).
XSLT is possibly the most horrible language to use for transforming documents. Languages such as CDuce are years ahead.
Rich.
I agree google analytics is helpful. However, it also increased my page load times by a long shot, so it had to be removed. It simply doesn't serve fast enough.
Did you try putting the Javascript somewhere other than the <head>? Obviously that's the recommended place, but in fact most of the functionality still works if you bury the Javascript down as close to </body> as it will go, and that should have less effect on the effective[1] page load time.
Rich.
[1] By "effective" I mean the time until the browser can render the page for the user, rather than the total load time.
Rich.
Unless the web sites get indexed by Google, and zombies use specially chosen keywords to search for their latest encrypted instructions.
Could be slow. Better for the botnet herders to buy Google AdWords adverts on obscure keywords using a stolen credit card, or make use of Google's free $50 coupons when they offer them.
Rich.
I'm sure I'm the only person on Earth who ever has to do this with any photo so I guess I can completely understand why "people" never what to do this.
You're not the only person on earth who wants to do this with the GIMP. For work I take screenshots of web pages and have to annotate them (text, rectangles, arrows) to show people what's wrong. I'd love to do this with the GIMP (I already use the GIMP for cropping, etc.), but I have no idea.
GIMP user since the late 90s.
Rich.
I'll third here.
I use GIMP quite a bit, although I'm by no means any sort of expert.
Please add the ability to do circles, lines, arrows, rectangles and so on in an obvious way.
Just the other day I had a screenshot which I was manipulating in GIMP, and all I wanted to do was to add an arrow pointing to some specific feature (it was a screenshot of a Google search, and I wanted to add an arrow pointing to the search box). Can this be done? Probably. Can this be done by me who's been using GIMP since the late 90s? No.
Rich.
Murders don't kill people.
Actually, I think you'll find they do ... :-)
Rich.
Its stupid comments like that that give OSS a bad name. Perhaps he's running a business full time and hasn't got the time/ability to fix it himself or the money to pay someone. Come down out of Geek Cuckoo Land and smell the real world.
Yeah, and while we're at it perhaps Microsoft should just give away all their software to people who have no money to pay for it. ... Right.
Rich.
I didn't want to move to Firefox 1.5. It worked slower and ate far mre RAM on my machines. But alas, few exploits later, I updated.
Because Firefox 1.0 support ended the moment 1.5 was out.
No it didn't. Official support for which you never paid a penny may have ended, but there was nothing to stop you taking the code & fixing it yourself, or paying a programmer to do it for you.
Rich.
The conundrum is that [third party open source developers, eg. Debian, are] trustworthy... right up until they screw you over.
Well you don't need to trust them. If you've got the source you can just look at the source and the patch (and even the vulnerability if it was a full-disclosure list) and check it for yourself. Or if you're not a competent programmer, pay a programmer on your behalf to do the check.
In many ways its funny to see the Windows closed-source-is-best Microsoft-is-always-right "community" acting this way. They can't see that it's obviously better in this case to have source, and third-party patches (with source) are clearly better for everyone.
Rich.
Rich.
My only defense is that the bubble hadn't even come to Texas yet, and nobody thought stock was worth the risk of working for a little startup.
Employees routinely get screwed even if they have good stock options. It's very common for the company - just before sale/IPO - to unilaterally change the terms of the deal so that employees are locked in for much longer than they agreed to be.
FWIW I once had options for 1% of a company which was later valued at (sticks pinky in mouth) 1 billllion dollars. I left before the IPO and before it was worth huge $$$, but the others I worked with got screwed in the above way and were only able to get out after the dot com boom had collapsed and the company was worth nothing.
Strangely enough the founders/directors didn't change their own terms and walked away with millions.
Rich.
Wow, 20 games which are exactly like all the other games around at the moment, but with slightly fancier graphics.
Where is the next ground-breaking playable game I can play with my friends, like Katamari Damacy or Super Monkey Ball or Mario Kart?
Rich.
2) Hastening the timely demise of AOL ...
Priceless
If only they'd released a 1.0 version that would never have happened...
A solution which forced us to regress to using Java or .Net is no solution at all BTW.
Rich.
Rich.
Back in 1994, Sun hired the core developer behind Tcl/Tk, and asked him to form a team around the language / graphical toolkit. The toolkit was very widely used and quite promising (for the time), but it languished at Sun and eventually they dumped it.
Rich.
So while I firmly believe that copying software illegal for personal use is a minor civil infraction, like speeding, and should be punished accordingly (a small fine that's enough to make you not want to do it but proportional to the harm) I believe that commercial copyright infringement is much more serious.
Where to start? Speeding endangers not just yourself but other road users and pedestrians. The amount of energy in your car increases with the square of speed making stopping harder and your car more deadly when it hits someone. Rules about speeding are one of the ways to prevent injury and death (rules against careless or dangerous driving, and driving under the influence being others). In the UK speeding is very much not a minor crime. People can be and are routinely banned from driving and imprisoned for speeding.
The "war on copying" is something like the modern "war on drugs". Computers are all about copying - it's their most basic operation. Networks have brought the price of copying down to almost nothing. The question becomes, in a world where distribution costs nothing, how to make sure people keep creating stuff? Should that be through the old industries who controlled distribution very tightly? Well, probably not. In the arts, it turned out that those industries were both very bad at producing works of art, and they didn't compensate the artists very well either. In software it seems that for whatever reason people will write the software for free and live by selling support.
Now I'm no more going to defend this guy who was caught selling millions of $$$s of software illegally. He's rather like a crack dealer. I may think that coca-derived products should be legalised (ie. regulated), but I won't defend someone who makes his living in the current unregulated environment selling crack to anyone.
Rich.