I didn't say "effective" though, did I? I said useful. I'd guess that most people want to recieve email from people they haven't whitelisted, making whitelists far, far less than perfect.
Why not use something like gnupg to sign email in order to prove the identity of the sender?
Because that requires changes to end-user behaviour.
In any case, a signature doesn't prove identity unless you or somebody in your web of trust has checked the fingerprint. This means that it's only a little more useful than a manual whitelist when it comes to avoiding spam.
No, it's well-known that humans make mistakes. Human decisions, when faced with hundreds of spam emails, result in false positives and false negatives as well. The comment you mention merely points out that they consider it to make less false negatives than the average human.
Outside of source with bizarre licenses, source that clearly furthers a Microsoft agenda (such as the Installer SDK Wix), etc..., does anything substantial exist?
Many people would call the GPL a "bizarre license".
I don't see why open-source furthering Microsoft's agenda makes something any less open-source. The whole point of the GPL is to further the Free Software agenda.
As to the consistency thing - the point was that almost all Windows apps, and basically all OS X apps use the same UI and widgets and whatnot.
What? I seem to remember Office having those damn annoying menus that keep hiding options before the rest of the OS, and something similar happened with Internet Explorer. Third-party applications often don't follow the normal UI, and sometimes Microsoft throws normal UI completely out of the window with things like Pinball and Media Player.
The same applies to Mac OS X, only less so. For instance, some applications use that brushed metal look, and some don't.
nVidia has this feature in their drivers. you can also use the PowerToys (free) from Microsoft to add multiple desktops too.
Oh, I'm sure you can bolt enough third-party add-ons and power user utilities onto Windows to get it working acceptably in terms of the user-interface, but how many users are willing to go to the trouble of hunting down all sorts of different add-ons to supply what comes as standard with a desktop Linux distribution?
Ah, one more thing. ACLs, Encryption and Compression is available per default in WinXP Pro. They are advanced and is very experimental in Linux still.
That's nothing end-users should have to care about, but yes, the security model Windows uses is better than the traditional Unix security model.
Wake me when users can create their own Samba shares and set custom passwords and usernames for it and still be bound to the files and security of their own profile in the local system.
I'm not sure what you mean by "bound to the files...", but you can set up a share by right-clicking on a folder and selecting "Share".
Not sure what you mean by opening a directory over SSH.
I mean opening a directory on a remote computer as if it were a local directory, without needing anything other than SSH installed on the remote computer. No commands, third-party applications or anything, just a normal file manager window.
Also, if a user has a problem then he can just send a remote desktop help to his best buddy.
Yep, Linux has this too - Preferences | Internet & Network | Desktop Sharing.
"emerge mozilla-firefox" works fine and easy. But the user needs to restart his gnome-panel for the new menuitems to show!
I don't think you can count problems in the power-user, bleeding-edge, source-based distribution Gentoo as problems with distributions aimed at normal users. Do normal distributions have this problem? All I know is that everything works as expected for me (using Gentoo, actually, but that's beside the point).
The user buys a new shiny USB mouse with 4 buttons and a scroll wheel to replace the 2 button one. Don't come and say it is "easy" to edit the xf86config file to fix the scrolling and the new buttons! At least in Windows it would work directly when you plug it in.
Whenever I've used a normal-user-aimed distribution like Mandrake in the past few years, it's detected my mouse properly. I don't consider editing config files to set up a mouse to be user-friendly, so don't try and put those words in my mouth.
BTW, LiteStep, WindowMaker, ThemeXP and many exists for windows to give whatever look you want. There are even OSX themes (including the dock with smooth hover effects, shadows etc too).
...and yet on Linux, the attitude is "users get confused by the availability of multiple interfaces"? Sounds like double-standards to me.
Linux is great. I use it everywhere I can. But, for it to gain consumer acceptance it really really needs a better and much more intuitive interface.
I agree. However, I think this because I think that Linux has to be substantially better than the competition to get consumer acceptance. I do not think this because I think Linux is far behind the competition - if I did, I wouldn't have been using it as my main desktop machine for the past five or six years. For instance, to make inroads on Windows' market share, it needs an impeccable installer that can cope with multiple operating systems sharing the same computer. Windows, on the other hand, just needs to carry on being installed by OEMs.
It also comes in multiple flavours and is not as compatible with as much software and hardware as previous versions. Plenty of people are picking Windows 2000 over Windows XP.
If you have a hard time deciding between new and old versions of the SAME operating system, bye the SAME vendor then you have larger problems.
If you count not blindly picking a suboptimal choice just because I think "newer is better" as a problem, then yes, I have a problem.
The fact you even tried this argument tells me you are part of the linux desktop problem and not part of the solution. Making rediculous comparisons like this does more for your ego than any OSS movement. It's called being ignorant my friend.
I am ignorant, a problem-causer and have a big ego, just because I dared point out that there are multiple versions of Windows to choose from? That's a pretty big leap of logic.
"newbie" is an elitist term, try calling them "users" instead. You'll go much farther without the elitism.
"Newbie" is simply shorthand for "new user", and I don't consider it pejorative. You'll go much farther without making rash assumptions about a person's attitude based upon the jargon they use.
Hence the problem, UI inconsistency. Or did you just skip that paragraph in my post?
What paragraph? You buried "unified interfaces are better" in the middle of a paragraph about "looking like Windows" (sounds like consistency to me) and lack of innovation.
Yeah, I missed those four words to begin with, but Redhat was making GNOME and KDE work alike a couple of releases ago, and since then things like GTK-Qt have come about.
Virtual desktops is a horrible way to manage windows which is why Apple created Expose and why MS is implimenting similar tech into Longhorn.
That's funny, users have been happily using them for years. And, in comparison to Windows, it's far, far better than *nothing*.
Wake me up when the UNIX developers design something more intelligent than virtual desktops, which by the way was a long long time ago. I ask, what have they done since? Nothing.
At the risk of repeating myself, what about Looking Glass?
First of all, which Linux? Redhat, slack, debian, Mandrake or maybe SuSE? That alone is reason enough for a customer to just pass it by and pick up a copy of windows.
Which Windows? Windows ME? Windows 2000? Windows XP Home? Windows XP Professional?
Oh then you need to pick a desktop, KDE, gnome, Windowmaker, etc etc etc etc.
I was under the impression most newbie-friendly distros had a clear default.
Oh but the apps I want to use are for gnome and my Linux install is using KDE.
What's the problem? You can run GNOME apps under KDE and vice-versa.
I see ZERO innovation in interface design from the Linux folks.
Wake me up when Windows gets multiple desktops, a feature supported by competing operating systems for at least a decade. Or when you can just open a directory on another machine via SSH. What about the recent LookingGlass beta? Does that not count as innovation? What has Microsoft released that is similar to that?
No, elements are usually block-level elements, so that is why you are able to suggest a width or height for them.
Re:People still use a shell for Linux?
on
Bash 3.0 Released
·
· Score: 1
The point of a shell script...
We've been talking about interactive use, not scripting. If I was going to script something, I'd use Python or Perl, not a shell script. Shell scripts are pretty hideous compared with any decent scripting language. "The right tool for the job" and all that...
Re:People still use a shell for Linux?
on
Bash 3.0 Released
·
· Score: 1
Starting up that explorer, selecting the files (manually checking if I have all of them, and only the right ones, perhaps, or having to scroll), then doing right click, is slower for me than typing in that command line.
I daresay it is. That's because you aren't comparing similar tasks any more. You are missing out starting up a shell, switching to the right directory, selecting a filename pattern that matches the files you want, etc. The previous command-line example assumed that you were logged in and had already chosen the files.
Besides, Konqueror offers shortcuts - if you think it's easier to type in a filename pattern, you can do so and have only those files showing.
Besides, I can still do the command line thing logged in over ssh
What's your point? X is network transparent too, and Konqueror can perform the same tasks over ssh, ftp, and a range of other protocols even if X wasn't network transparent.
Re:People still use a shell for Linux?
on
Bash 3.0 Released
·
· Score: 1
You misunderstand. The right-click... action only has to be performed once, and it operates on all selected files. You don't have to perform the action a thousand times.
Re:People still use a shell for Linux?
on
Bash 3.0 Released
·
· Score: 2, Interesting
Bad example. Using KDE, I click on my home directory icon, select the images I want to convert, right-click on one of them and pick Actions | Convert To | PNG.
This is just as quick, doesn't require you to memorise complicated syntax, and doesn't require filenames that follow a common pattern.
Cautionary tale? Not at all. The developers know that their contributions may be used by companies. I see no difference to whether it's their ex-employer or other companies that do it. The developers aren't forced to continue contributing to Mozilla after AOL stops employing them to do that.
Obviously its not that hard to understand what I was talking about, cause you got it.
Well I had to re-read your post a few times to work it out, and I had to ask to be certain. It's a lot easier to just use the right words in the first place.
The point was not which browser was right, the point is that they differ.
Well the whole point of the specification is to get all the browsers agreeing on what stylesheets mean. If one browser ignores the specification, and the resulting mismatch causes you problems, then the fact that the browser is ignoring the specification is very relevent.
What is a non-replaced in line level element?
The section of the specification I referred you to had links to their definitions.
Lastly if someone comes to a web site and decides to spoof a useragent, then their browser better be able to render the content like the useragent they are spoofing.
You missed off the second part of my sentence, "User-Agent headers are routinely spoofed or missing".
However in the dom there are otherways of spoofing out a useragents capabilities. EG:
if ( document.getElementById ) { do your document.getElementByID }
A web browser cant fake that kind of sniffing.
Of course, and sniffing a visitor's capabilities is a good technique to use. But your initial response explicitly referred to detecting the user-agent. Those are two different ways of approaching the same problem.
I think he hit th enail on the head - how many times do you see someone looking for an OSS aka "free" counterpart to a CSS aka "cost money" product? They're looking for free as in no cost, not as in I can mod it. That perception will limit entry and ultimately stifle innovation. How many innovative, vs "let's copy the functionality of product X" OSs programs are out there?
You catch more bees with honey than vinegar. In my experience, most people are perfectly happy to copy proprietary software illegally, and are not capable of modifying software.
Threatening users by saying it's illegal to copy is vinegar. Competing on features is honey. The thing is, hardly anybody will switch away to software that does less. So if Free Software authors want to actually get people to switch, their first priority is the features that the proprietary software already has. This is why cloning is not only natural but beneficial.
Remember that proprietary software that is widely used is widely used for a reason - the features are actually useful. You expect Free Software authors to avoid useful features because proprietary software has them as well?
how does it go into that list noting it's absolutely not wanted?
If it has a quality value of 0 (application/xhtml+xml;q=0). In practical terms, it just has to have a lower quality value than */* (which, in a decent browser, would have a low quality value itself).
P means paragraph, and an image is not a paragraph. If they're willing to add a src attribute to p to make it insert images
You still aren't getting it. The src attribute is generic - it's not a new attribute for the <p> element type, it's a new attribute for most element types. It's a generic way of replacing existing content with external resources.
The example in question starts out like this:
<p>foo</p>
It's a paragraph of text. That is correct markup. Then, when the src attribute is added, the resource replaces that entire element if possible. The fact that the element happens to be a <p> element is irrelevent; it could be a <table> element, an <h1> element, or whatever. The element type with the src attribute isn't describing the resource pointed to by the src attribute, you are looking at it backwards.
It doesn't, that's not what is happening. There is a paragraph "foo", and if it is possible to replace it with the content indicated by the src attribute, it will be replaced.
You see, that's just the problem: firstly, that's far out of reach of the average web developer. But more importantly, the ones who know enough to get by, but not all the issues, will write code like this.
Problem number one: your code will incorrectly serve application/xhtml+xml to browsers that specifically say they don't want it under any circumstances.
Problem number two: your code will also favour application/xhtml+xml to browsers that support it, but explicitly say they prefer text/html (e.g. if they can get by, but have poor support compared with normal HTML).
Problem number three: in situations where caching takes place, your code can be responsible for serving application/xhtml+xml to browsers that can't handle it and don't mention it in their Accept header at all.
In the first and last cases, problems result in an utterly broken website. In the middle case, users may get a suboptimal experience compared with if you didn't try and serve application/xhtml+xml.
Read RFC 2616, and pay close attention to the Accept header and the Vary header.
What does XHTML support actually bring you? Is there any benefit, or are you just doing it because it's there?
Given that bugs like this are going to crop up as developers don't really know the edge cases in content negotiation or XHTML, is it worth advocating that people move to XHTML when there is virtually no benefit and plenty of places to trip up and cause severe problems?
The only version of XHTML that is suitable for transmission as text/html is XHTML 1.0 following Appendix C. XHTML transmitted in this fashion doesn't have any of the benefits of mixed namespaces, stricter parsing, etc.
You get these benefits when you serve XHTML as application/xhtml+xml, and your visitors use browsers that support those features (such visitors are extremely rare - SVG isn't even in main Mozilla builds yet). But many legacy user-agents require text/html. Search engines would probably be the most important ones.
So unless you are willing to set up content-negotiation, sending different document types to different browsers, and unless you have a niche market that use browsers that understand these new features, you really aren't going to get anything from XHTML. Not for a few years, anyway.
Something better would be to use pure XML for creating content and then let the browser apply a CSS to present the content.
No, that would be very much worse. The whole point of a publically specified XML application like XHTML is that everybody understand what the element types mean. If you go around inventing your own element types, nobody will no what you mean. Google understands <h1> as being more important than normal text. It won't understand <my-fancy-heading> in the same way.
Slashdot Mirror
I didn't say "effective" though, did I? I said useful. I'd guess that most people want to recieve email from people they haven't whitelisted, making whitelists far, far less than perfect.
Because that requires changes to end-user behaviour.
In any case, a signature doesn't prove identity unless you or somebody in your web of trust has checked the fingerprint. This means that it's only a little more useful than a manual whitelist when it comes to avoiding spam.
No, it's well-known that humans make mistakes. Human decisions, when faced with hundreds of spam emails, result in false positives and false negatives as well. The comment you mention merely points out that they consider it to make less false negatives than the average human.
Drop a file called userChrome.css into your chrome directory with the following contents:
You'll be able to find the right place by searching for a file called userChrome-example.css.
Many people would call the GPL a "bizarre license".
I don't see why open-source furthering Microsoft's agenda makes something any less open-source. The whole point of the GPL is to further the Free Software agenda.
Let's not have double standards here.
What? I seem to remember Office having those damn annoying menus that keep hiding options before the rest of the OS, and something similar happened with Internet Explorer. Third-party applications often don't follow the normal UI, and sometimes Microsoft throws normal UI completely out of the window with things like Pinball and Media Player.
The same applies to Mac OS X, only less so. For instance, some applications use that brushed metal look, and some don't.
Oh, I'm sure you can bolt enough third-party add-ons and power user utilities onto Windows to get it working acceptably in terms of the user-interface, but how many users are willing to go to the trouble of hunting down all sorts of different add-ons to supply what comes as standard with a desktop Linux distribution?
That's nothing end-users should have to care about, but yes, the security model Windows uses is better than the traditional Unix security model.
I'm not sure what you mean by "bound to the files...", but you can set up a share by right-clicking on a folder and selecting "Share".
I mean opening a directory on a remote computer as if it were a local directory, without needing anything other than SSH installed on the remote computer. No commands, third-party applications or anything, just a normal file manager window.
Yep, Linux has this too - Preferences | Internet & Network | Desktop Sharing.
I don't think you can count problems in the power-user, bleeding-edge, source-based distribution Gentoo as problems with distributions aimed at normal users. Do normal distributions have this problem? All I know is that everything works as expected for me (using Gentoo, actually, but that's beside the point).
Whenever I've used a normal-user-aimed distribution like Mandrake in the past few years, it's detected my mouse properly. I don't consider editing config files to set up a mouse to be user-friendly, so don't try and put those words in my mouth.
I agree. However, I think this because I think that Linux has to be substantially better than the competition to get consumer acceptance. I do not think this because I think Linux is far behind the competition - if I did, I wouldn't have been using it as my main desktop machine for the past five or six years. For instance, to make inroads on Windows' market share, it needs an impeccable installer that can cope with multiple operating systems sharing the same computer. Windows, on the other hand, just needs to carry on being installed by OEMs.
It also comes in multiple flavours and is not as compatible with as much software and hardware as previous versions. Plenty of people are picking Windows 2000 over Windows XP.
If you count not blindly picking a suboptimal choice just because I think "newer is better" as a problem, then yes, I have a problem.
I am ignorant, a problem-causer and have a big ego, just because I dared point out that there are multiple versions of Windows to choose from? That's a pretty big leap of logic.
"Newbie" is simply shorthand for "new user", and I don't consider it pejorative. You'll go much farther without making rash assumptions about a person's attitude based upon the jargon they use.
What paragraph? You buried "unified interfaces are better" in the middle of a paragraph about "looking like Windows" (sounds like consistency to me) and lack of innovation.
Yeah, I missed those four words to begin with, but Redhat was making GNOME and KDE work alike a couple of releases ago, and since then things like GTK-Qt have come about.
That's funny, users have been happily using them for years. And, in comparison to Windows, it's far, far better than *nothing*.
At the risk of repeating myself, what about Looking Glass?
Which Windows? Windows ME? Windows 2000? Windows XP Home? Windows XP Professional?
I was under the impression most newbie-friendly distros had a clear default.
What's the problem? You can run GNOME apps under KDE and vice-versa.
Wake me up when Windows gets multiple desktops, a feature supported by competing operating systems for at least a decade. Or when you can just open a directory on another machine via SSH. What about the recent LookingGlass beta? Does that not count as innovation? What has Microsoft released that is similar to that?
No, elements are usually block-level elements, so that is why you are able to suggest a width or height for them.
We've been talking about interactive use, not scripting. If I was going to script something, I'd use Python or Perl, not a shell script. Shell scripts are pretty hideous compared with any decent scripting language. "The right tool for the job" and all that...
I daresay it is. That's because you aren't comparing similar tasks any more. You are missing out starting up a shell, switching to the right directory, selecting a filename pattern that matches the files you want, etc. The previous command-line example assumed that you were logged in and had already chosen the files.
Besides, Konqueror offers shortcuts - if you think it's easier to type in a filename pattern, you can do so and have only those files showing.
What's your point? X is network transparent too, and Konqueror can perform the same tasks over ssh, ftp, and a range of other protocols even if X wasn't network transparent.
You misunderstand. The right-click... action only has to be performed once, and it operates on all selected files. You don't have to perform the action a thousand times.
Bad example. Using KDE, I click on my home directory icon, select the images I want to convert, right-click on one of them and pick Actions | Convert To | PNG.
This is just as quick, doesn't require you to memorise complicated syntax, and doesn't require filenames that follow a common pattern.
Cautionary tale? Not at all. The developers know that their contributions may be used by companies. I see no difference to whether it's their ex-employer or other companies that do it. The developers aren't forced to continue contributing to Mozilla after AOL stops employing them to do that.
Yes it does.
Well I had to re-read your post a few times to work it out, and I had to ask to be certain. It's a lot easier to just use the right words in the first place.
Well the whole point of the specification is to get all the browsers agreeing on what stylesheets mean. If one browser ignores the specification, and the resulting mismatch causes you problems, then the fact that the browser is ignoring the specification is very relevent.
The section of the specification I referred you to had links to their definitions.
You missed off the second part of my sentence, "User-Agent headers are routinely spoofed or missing ".
Of course, and sniffing a visitor's capabilities is a good technique to use. But your initial response explicitly referred to detecting the user-agent. Those are two different ways of approaching the same problem.
You catch more bees with honey than vinegar. In my experience, most people are perfectly happy to copy proprietary software illegally, and are not capable of modifying software.
Threatening users by saying it's illegal to copy is vinegar. Competing on features is honey. The thing is, hardly anybody will switch away to software that does less. So if Free Software authors want to actually get people to switch, their first priority is the features that the proprietary software already has. This is why cloning is not only natural but beneficial.
Remember that proprietary software that is widely used is widely used for a reason - the features are actually useful. You expect Free Software authors to avoid useful features because proprietary software has them as well?
If it has a quality value of 0 (application/xhtml+xml;q=0). In practical terms, it just has to have a lower quality value than */* (which, in a decent browser, would have a low quality value itself).
You still aren't getting it. The src attribute is generic - it's not a new attribute for the <p> element type, it's a new attribute for most element types. It's a generic way of replacing existing content with external resources.
The example in question starts out like this:
<p>foo</p>It's a paragraph of text. That is correct markup. Then, when the src attribute is added, the resource replaces that entire element if possible. The fact that the element happens to be a <p> element is irrelevent; it could be a <table> element, an <h1> element, or whatever. The element type with the src attribute isn't describing the resource pointed to by the src attribute, you are looking at it backwards.
It doesn't, that's not what is happening. There is a paragraph "foo", and if it is possible to replace it with the content indicated by the src attribute, it will be replaced.
You see, that's just the problem: firstly, that's far out of reach of the average web developer. But more importantly, the ones who know enough to get by, but not all the issues, will write code like this.
Problem number one: your code will incorrectly serve application/xhtml+xml to browsers that specifically say they don't want it under any circumstances.
Problem number two: your code will also favour application/xhtml+xml to browsers that support it, but explicitly say they prefer text/html (e.g. if they can get by, but have poor support compared with normal HTML).
Problem number three: in situations where caching takes place, your code can be responsible for serving application/xhtml+xml to browsers that can't handle it and don't mention it in their Accept header at all.
In the first and last cases, problems result in an utterly broken website. In the middle case, users may get a suboptimal experience compared with if you didn't try and serve application/xhtml+xml.
Read RFC 2616, and pay close attention to the Accept header and the Vary header.
What does XHTML support actually bring you? Is there any benefit, or are you just doing it because it's there?
Given that bugs like this are going to crop up as developers don't really know the edge cases in content negotiation or XHTML, is it worth advocating that people move to XHTML when there is virtually no benefit and plenty of places to trip up and cause severe problems?
The examples you gave are not equivalent. The equivalent of: />
...is:
<img src="foo.png" alt="foo"
<p src="foo.png">foo</p>
The complicated example you gave is merely an example of how you can provide multiple fallbacks.
The only version of XHTML that is suitable for transmission as text/html is XHTML 1.0 following Appendix C. XHTML transmitted in this fashion doesn't have any of the benefits of mixed namespaces, stricter parsing, etc.
You get these benefits when you serve XHTML as application/xhtml+xml, and your visitors use browsers that support those features (such visitors are extremely rare - SVG isn't even in main Mozilla builds yet). But many legacy user-agents require text/html. Search engines would probably be the most important ones.
So unless you are willing to set up content-negotiation, sending different document types to different browsers, and unless you have a niche market that use browsers that understand these new features, you really aren't going to get anything from XHTML. Not for a few years, anyway.
No, that would be very much worse. The whole point of a publically specified XML application like XHTML is that everybody understand what the element types mean. If you go around inventing your own element types, nobody will no what you mean. Google understands <h1> as being more important than normal text. It won't understand <my-fancy-heading> in the same way.