“Proper quote marks” I think — and also proper long dashes — make the web a more beautiful place. You don't even need any special software to use them since most programs and web services now use Unicode and just allow you to type the character and then display, store, and transmit it correctly.
More nice typography: CSS 3.0 will include some hyphenation support and for those who can't wait there is already a nice bookmarklet that does hyphenation on the fly in any modern browser! If you ever have to include a very long word (such as Dampfschiffahrtsgesellschaft oder Gesellschafterversammlung) in a text that might end up in a narrow column (such as next to an image) or simply on a mobile device, you can already easily save your layout by inserting a soft hyphen (HTML entity ). This will tell any modern web browser where to break the word. Here's the example used by the W3C: antiestablishmentarianism, antiestablishmentarianism, and antiestablishmentarianism. Resize your window (or zoom) to see how your web browser automatically hyphenates the word! (No Javascript required.)
Further Reading:
Appendix: Soft hyphen support in browsers and search engines
Internet Explorer before version 5.0 always displays soft hypens. That's terrible, but IE version ≤ 5 are virtually extinct now, so the problem is less. IE ≥5.0, Firefox ≥3.0, Opera ≥9.0.2, Safari ≥2, and even Lynx ≥2.8.2 all treat soft hyphens correctly, while those other browsers who don't support them will just ignore them. (Source: [2])
Google and other major search engines correctly ignore soft hyphens. (Same source)
In other words: soft hyphens are safe and beneficial to use as of now.
Showing posts with label computers. Show all posts
Showing posts with label computers. Show all posts
28 September 2009
28 June 2009
GMail and SPAM from my professor
I am using GMail since 2005 and for the last four years it did an awesome job of keeping me clear from SPAM. I've had months where not a single SPAM message was delivered. Now, suddenly, I am regularly finding some SPAM in my inbox, most of which has the email address from one of my professors set as the sender. At first I thought, he might have caught a virus that's abusing his computer, but that would have stopped by now, so I think that the spammers are just faking his email address. Oddly enough, the university's spam filter marks the message as SPAM, but GMail doesn't filter it out.
I think here's why: the SPF sender authentificationn declares the sender as permitted, so GMail (rightly) assumes that the professor is really the sender, therefore the mail will not go to the SPAM folder. I think this is right: email from somebody who has sent me many non-spam messages and whom I even sent some messages back should always be trusted. The problem seems to be that some spammer has managed to send SPAM from within inside the university.
I will ask our admins to look into this.
I think here's why: the SPF sender authentificationn declares the sender as permitted, so GMail (rightly) assumes that the professor is really the sender, therefore the mail will not go to the SPAM folder. I think this is right: email from somebody who has sent me many non-spam messages and whom I even sent some messages back should always be trusted. The problem seems to be that some spammer has managed to send SPAM from within inside the university.
I will ask our admins to look into this.
24 June 2009
firefox vs web applications, part III
So after all the trouble my idea from the last post gave me, after seeing how Firefox could not really handle different menu/nav/status bar settings for different windows, because new windows would always get it wrong, I thought why can't we just get rid of that menu bar for all windows. Internet Explorer and Google Chrome don't have a menu bar either! Now, how other browsers look is clearly not an argument for me to change my browser, but I know that a lot of other people like to copy things and so I thought: "There must be a Firefox Add-on that gives Firefox a Google-Chrome-like skin." I've used any skins myself, but I know it's a big thing and if it can help me, let's look for it! And indeed! Even on the first page of recommended Firefox Add-Ons I found one that's named "Personal Menu" and it's prime feature is "to disable the menu bar". I installed it and it's great! It offers a few extra buttons for the nav bar (like a bookmarks and a history button) plus one button that can be entirely customized to contain just those few menu items that one personally needs.
Here's how my Firefox looks like now:
I am keeping the location box and the search box, because I use both often. I like how Firefox now manages bookmarks and history from the location bar!
For comparison, on Windows my Firefox still looks as before:
As you can see, I saved some space by getting rid of the refresh/stop/home buttons which I all control by keyboard shortcuts. Actually I also control the back/forward function by keyboard but I felt that a web browser just doesn't look like a web browser without those buttons. Also, I switched to "small buttons".
Now, I saved one line of screen real estate and I like it! On many pages I still need to do F11 (full screen) to read them well, but I think the browser has become better with much effort. That's good!
Here's how my Firefox looks like now:
For comparison, on Windows my Firefox still looks as before:
Now, I saved one line of screen real estate and I like it! On many pages I still need to do F11 (full screen) to read them well, but I think the browser has become better with much effort. That's good!
19 June 2009
Firefox vs web applications part II (notes)
apparently the only way to remove the menubar from a window is to open a new window via javascript (JS) window.open(url, window_name, 'menubar=no'). This code does not work from a bookmarklet (at least not naively), so I wrote a simple "web application launcher page" which contains links to my web applications via JS.
The syntax of the above JS function also supports a parameter 'status=no' to switch of the status-bar, but this is ignored by Firefox. Actually, whatever the parameters for the new window are, Firefox will recognize the "intention" of opening a new window and will then do it using its own parameters: the location bar will always be visible, but grayed-out (read-only, which is very sensible in general, but not so much for my applications), and the status bar will be just as in the current setting (menu) "View - Status bar". Since the menu is disabled in the new window, I have to disable the status bar manually in my "web application launcher" window and then click on the link, which gives me a new window without status bar. The same is true for the location bar. I have to disable it manually before I click on the JS link. This is more complicated than it should be!
Best would really be to let me disable the menubar manually, too, and then have the browser simply remember the setting of status/menu/location bar for each website where I actively changed it. Being at it once, it should also remember the window size and position for each website where I manually change it.
There is also a terribly nasty bug: When I open Facebook in this way without a menu bar (no matter what the settings for status and location bar), it will open without scroll bars, but GMail with the same settings has scrollbars! Fortunately I could fix it by adding "scrollbars=yes" to the window.open() call. Just perplexing that GMail does not need it.
Next step for me is to try living with my javascript launcher page and see if it does the job. Stay tuned and tell me if you find alternative solutions.
The syntax of the above JS function also supports a parameter 'status=no' to switch of the status-bar, but this is ignored by Firefox. Actually, whatever the parameters for the new window are, Firefox will recognize the "intention" of opening a new window and will then do it using its own parameters: the location bar will always be visible, but grayed-out (read-only, which is very sensible in general, but not so much for my applications), and the status bar will be just as in the current setting (menu) "View - Status bar". Since the menu is disabled in the new window, I have to disable the status bar manually in my "web application launcher" window and then click on the link, which gives me a new window without status bar. The same is true for the location bar. I have to disable it manually before I click on the JS link. This is more complicated than it should be!
Best would really be to let me disable the menubar manually, too, and then have the browser simply remember the setting of status/menu/location bar for each website where I actively changed it. Being at it once, it should also remember the window size and position for each website where I manually change it.
There is also a terribly nasty bug: When I open Facebook in this way without a menu bar (no matter what the settings for status and location bar), it will open without scroll bars, but GMail with the same settings has scrollbars! Fortunately I could fix it by adding "scrollbars=yes" to the window.open() call. Just perplexing that GMail does not need it.
Next step for me is to try living with my javascript launcher page and see if it does the job. Stay tuned and tell me if you find alternative solutions.
15 June 2009
web applications and Firefox
I am currently using the following web applications:
Wouldn't it be great if Firefox had a web application mode, just like Google Chrome, where the menu and location bar are not displayed? Given that the navigation bar can already be hidden and shown easily, all the is necessary would be to save the visibility state of the bar per each website! So easy to implement!
I just read about the Mozilla Prism Project which is a Browser specific to web applications, but it doesn't seem to be very popular. Their is also a Prism extension for Firefox which I just tried out, but it just doesn't work. The only way to start a webapplication is to first create a shortcut on the desktop (why not just a bookmark?!). I did that and clicked the shortcut only to be left with an ordinary and empty Firefox window. Sucks!
Currently I am experimenting a little and run my Firefox without any location bar. It's actually not too bad, because I can still use Ctrl-L to type adresses. It's the same behavior only using a little dialog window that disappears after hitting Return. Let's see if that can take away some pressure until the Firefox developers show reason.
- GMail
- Google Docs
- Google Reader
Wouldn't it be great if Firefox had a web application mode, just like Google Chrome, where the menu and location bar are not displayed? Given that the navigation bar can already be hidden and shown easily, all the is necessary would be to save the visibility state of the bar per each website! So easy to implement!
I just read about the Mozilla Prism Project which is a Browser specific to web applications, but it doesn't seem to be very popular. Their is also a Prism extension for Firefox which I just tried out, but it just doesn't work. The only way to start a webapplication is to first create a shortcut on the desktop (why not just a bookmark?!). I did that and clicked the shortcut only to be left with an ordinary and empty Firefox window. Sucks!
Currently I am experimenting a little and run my Firefox without any location bar. It's actually not too bad, because I can still use Ctrl-L to type adresses. It's the same behavior only using a little dialog window that disappears after hitting Return. Let's see if that can take away some pressure until the Firefox developers show reason.
14 June 2009
crazy linux: so easy and so complicated
Linux-Fans would probably claim it as an advantage of their OS that you can easily move an installation from one hard-disk or partition to another, only need to update boot-manager and fstab, can even keep all your directory-names.
In reality, this simple maneuver can easily lead to a big mess (as in: the system does not boot any more and gives loads of contradictory messages). Here's what happened to me.
I wanted to update my Ubuntu from 8.04 to 8.10 (and maybe 9.x later). One reason to do it is that apparently I can then install Facebook-chat in Pidgin. But this update needs 1.5 Gigs of free space on the root partition and so I had to make a larger partition and move my installation there.
First, it was really easy to create a new partition. Before formatting it, however, I had to restart the computer, just as I would have needed with Windows. Formatting it, was also easy, just call "mkfs.ext3" which I don't remember exactly because it pops up conveniently when pressing TAB at the shell. (Yes, I did partitioning and formatting by hand, because there was no graphical tool for it installed by default and finding one to install would have taken longer...)
Second, I googled on how to copy files, read something about "cp -a" and tried it out. That was a big mistake. The correct command would have been "cp -ax" where the "-x" makes sure that file systems which are mounted under / are not also copied. I was not only so unlucky to waste a lot of time by copying useless stuff (had my windows partition mounted...), it also apparently crashed my computer when it tried to copy some of the virtual file systems under /sys.
After this crash, I copied the remaining data individually (so as to not copy things twice). Since all is working now, I think I didn't miss anything.
Third, making the new partition the one being used is really as easy as changing GRUB's menu.lst and /etc/fstab, and it did it right on the third try, but I still do not know at all what when wrong on my previous tries. On the first try, I just walked into menu.lst and replaced the kernel's root= parameter with the new partition name sda7, ignoring all the fancy UUID names for other partitions which I found very confusing. (And still don't get the point... it makes the system transparent for hard-disk switches, but makes it vulnerable to hard-disk reformatting...).
This first try actually booted my system from the new partition (yay!), but when I typed 'mount' at the prompt, it would show the old partition to be mounted on /. But it actually was the new one. What a blatant lie! This is a severe error on 'mount's side.
Anyway, I decided to do the update of 'fstab', hoping this would solve the problem. I added a line mounting the new partition as /, and changed the line for the old partition to mount as /part1. All obvious things you would think, right? But this actually messed up my boot and brought me more blatant lies. After restarting, there was a weird text-mode message box signed by 'gdm' about X not being able to start. Part of the box was overwritten by a text-mode logon prompt. I logged in and had 'mount' lie to me again. It told me that no partition but the former / was mounted when in fact everything else was mounted. Sucker!
I booted back into the old system (which still worked), google for UUID and then changed the 'menu.lst' as well as 'fstab' to use the UUID instead of /dev/sda7. I also corrected a mistake in fstab where I had typed "default" instead of "defaults". Anyway, in my opinion none of those things should have made a difference, but in reality the new system booted thereafter and I couldn't find any problems any more. Case considered close. Reason for trouble unknown.
It's weird how my perception of computers has changed. It used to be that I wanted to understand how things work, usually helped by reading the manual and trying different things out, and then from that understanding knew how to achieve what I wanted. But now, understanding is totally absent from the process. I just assume that everything will somehow work and when I run into trouble (and only then), I ask Google and try out what somebody else wrote, not even caring who posted it and where.
This is really a black vision of computing; one that's totally opposed to my research approach: in verified programming and refinement, everything is meant to be transparent to people and every step understandable. I really sucks that computers in practice have become like black boxes that we can only control by trial and error, not logic or understanding.
After all, it was easy to copy the Ubuntu installation onto a new partition and I am now happily downloading my update. But the way it went is just not right.
Edit: I had the update run overnight because the download rate was so slow (that's GradHouse...) and it bothered me in the early morning with some decisions that pop up randomly during the update. Overall, Ubuntu now uses 0.9 Gigs more of hard disk space, has a new "switch computer off or change chat status" button, needed reconfiguration of my Quickstart-buttons (those extra keys on the laptop-keyboard), because their keycodes apparently got renamed... Most other things still seem to work and most importantly: I could install Facebook-chat for Pidgin from Google-Code and that works, too. Yay!
In reality, this simple maneuver can easily lead to a big mess (as in: the system does not boot any more and gives loads of contradictory messages). Here's what happened to me.
I wanted to update my Ubuntu from 8.04 to 8.10 (and maybe 9.x later). One reason to do it is that apparently I can then install Facebook-chat in Pidgin. But this update needs 1.5 Gigs of free space on the root partition and so I had to make a larger partition and move my installation there.
First, it was really easy to create a new partition. Before formatting it, however, I had to restart the computer, just as I would have needed with Windows. Formatting it, was also easy, just call "mkfs.ext3" which I don't remember exactly because it pops up conveniently when pressing TAB at the shell. (Yes, I did partitioning and formatting by hand, because there was no graphical tool for it installed by default and finding one to install would have taken longer...)
Second, I googled on how to copy files, read something about "cp -a" and tried it out. That was a big mistake. The correct command would have been "cp -ax" where the "-x" makes sure that file systems which are mounted under / are not also copied. I was not only so unlucky to waste a lot of time by copying useless stuff (had my windows partition mounted...), it also apparently crashed my computer when it tried to copy some of the virtual file systems under /sys.
After this crash, I copied the remaining data individually (so as to not copy things twice). Since all is working now, I think I didn't miss anything.
Third, making the new partition the one being used is really as easy as changing GRUB's menu.lst and /etc/fstab, and it did it right on the third try, but I still do not know at all what when wrong on my previous tries. On the first try, I just walked into menu.lst and replaced the kernel's root= parameter with the new partition name sda7, ignoring all the fancy UUID names for other partitions which I found very confusing. (And still don't get the point... it makes the system transparent for hard-disk switches, but makes it vulnerable to hard-disk reformatting...).
This first try actually booted my system from the new partition (yay!), but when I typed 'mount' at the prompt, it would show the old partition to be mounted on /. But it actually was the new one. What a blatant lie! This is a severe error on 'mount's side.
Anyway, I decided to do the update of 'fstab', hoping this would solve the problem. I added a line mounting the new partition as /, and changed the line for the old partition to mount as /part1. All obvious things you would think, right? But this actually messed up my boot and brought me more blatant lies. After restarting, there was a weird text-mode message box signed by 'gdm' about X not being able to start. Part of the box was overwritten by a text-mode logon prompt. I logged in and had 'mount' lie to me again. It told me that no partition but the former / was mounted when in fact everything else was mounted. Sucker!
I booted back into the old system (which still worked), google for UUID and then changed the 'menu.lst' as well as 'fstab' to use the UUID instead of /dev/sda7. I also corrected a mistake in fstab where I had typed "default" instead of "defaults". Anyway, in my opinion none of those things should have made a difference, but in reality the new system booted thereafter and I couldn't find any problems any more. Case considered close. Reason for trouble unknown.
It's weird how my perception of computers has changed. It used to be that I wanted to understand how things work, usually helped by reading the manual and trying different things out, and then from that understanding knew how to achieve what I wanted. But now, understanding is totally absent from the process. I just assume that everything will somehow work and when I run into trouble (and only then), I ask Google and try out what somebody else wrote, not even caring who posted it and where.
This is really a black vision of computing; one that's totally opposed to my research approach: in verified programming and refinement, everything is meant to be transparent to people and every step understandable. I really sucks that computers in practice have become like black boxes that we can only control by trial and error, not logic or understanding.
After all, it was easy to copy the Ubuntu installation onto a new partition and I am now happily downloading my update. But the way it went is just not right.
Edit: I had the update run overnight because the download rate was so slow (that's GradHouse...) and it bothered me in the early morning with some decisions that pop up randomly during the update. Overall, Ubuntu now uses 0.9 Gigs more of hard disk space, has a new "switch computer off or change chat status" button, needed reconfiguration of my Quickstart-buttons (those extra keys on the laptop-keyboard), because their keycodes apparently got renamed... Most other things still seem to work and most importantly: I could install Facebook-chat for Pidgin from Google-Code and that works, too. Yay!
Subscribe to:
Posts (Atom)