Captured Moments » Eventually interesting stuff about Linux, Programming, Software, Photography.

Tags

  PHP     Photography     Nikon     Photoblog     Plugin     Programming     Usability     GPS     Weather     Hugin     Live     C++     Work     Sony     F-Spot     Qt     Panorama     Web     Windows     UFRaw         Software     Gimp     Linux     Ubuntu     Bibble  

My Flickr Photos

Books on blurb.com

aus der ev. Kirche Sulzbach (Ts.)
Julia und Paul
Wellner Bou

Admin area

I am working on a desktop without antialiased fonts, as this is much sharper and clearer using small font sizes as I am doing. As desktop font I am using Microsoft’s Tahoma, as I think this is the most space saving and at the same time best looking font without using antialiasing. (If you know another one, please tell me!)

antialias.png

If you do and you are using Mozilla Firefox and Thunderbird, you probably will have noticed that the new versions (Firefox 3.5 and Thunderbird 3) always use antialliased fonts. Even trying to change the setting

gfx.use_text_smoothing_setting
in the about:conf page does not change anything.

This bug is reported at Ubuntu’s Launchpad. I crawled through the comments and workarounds given there, and the solution which worked great for me without breaking any other stuff was the solution mentioned in comment #52. Save this as ~/.fonts.conf:

<?xml version="1.0"?>
<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
<fontconfig>
 <match target="font" >
  <edit mode="assign" name="rgba" >
   <const>rgb</const>
  </edit>
 </match>
 <match target="font" >
  <edit mode="assign" name="hinting" >
	
   <bool>true</bool>
  </edit>
 </match>
 <match target="font" >
  <edit mode="assign" name="hintstyle" >
   <const>hintfull</const>
  </edit>
 </match>
	
 <match target="font" >
  <edit mode="assign" name="antialias" >
   <bool>true</bool>
  </edit>
 </match>
</fontconfig>

Restart Thunderbird or Firefox.

Aside from the fact that there is again a discussion around the time bug I updated my branch fixing this feature on gitorious.

To use it, just grap the sources and compile it.

git clone git://gitorious.org/~paulwb/f-spot/paulwbs-clone.git

As an alternative you can download the tarball linked on the gitorious page.

cd paulwbs-clone/
git checkout origin/ImportTimeHandling
./autogen.sh
make
sudo make install

Looking forward to have this fixed and working.

Update: Steve created a PPA for Ubuntu users which fixes this bug just as my branch is doing. And, for more information, the Ubuntu bug is here.

The most common spam blocking technique for web forms used nowadays are captchas. I think this is the most effective way, too, as it is difficult to process and read the captcha images automatically and it is quite expensive in terms of CPU and memory resources.

There are a lot of alternative way out there, more or less brilliant, more or less useful. All of those anti spam measures I know not using captchas would not work if they would find a wide usage, as with a little bit more of spam bot intelligence they would be useless.

  • Referrer check: A simple one, known to not work as this can be (and is actually) faked easily in automatic HTTP requests.
  • Using a token as described here years ago using the user session is easy to break implementing sessions and cookies into the spam bot.
  • A mathematical question: This is used in Flatpress’ accessible antispam plugin and there is an extension for Typo3 implementing this, too. It is not widely used in comparison to captchas and this might be the reason that it might work for some cases. It would be easy to break.

Especially the last one is, in my eyes, nonsense. It requires user interaction where there is absolutely no need to. This is no turing test, this is a simple calculation test. And who is faster in simple calculations than computer programs? The only problem for a computer program would be to read the problem. Of course, this could be made difficult hiding it in a separate element somewhere in the html document, using CSS and JavaScript to display it to the human user where it should appear. Although this may reduce the accessibility.

There is another way to achieve the same level of spam protection making it significantly more user friendly:

<form name="commentform">
    <label for="intput_a">Name:</label><input name="intput_a" type="text"/>
    <label for="intput_b">Email:</label><input name="intput_b" type="text"/>
    <label for="intput_c">Web:</label><input name="intput_c" type="text"/>
    <label for="intput_d">Leave this field empty!</label><input name="intput_d" type="text"/>
    <label for="intput_e">Comment:</label><textarea name="intput_e"></textarea>
</form>

Know what I mean? Why asking the user to put something and why not asking him to explicitly put nothing?

You could use CSS to hide the label and input which should be left empty so that a normal user won’t be bothered. You could even leave the label text out and hide the input with CSS if you assume that all user agents visiting your site support a basic kind of CSS.

Spam bots will try to fill something in this field and so you can recognize easily the spam entries. I use this method on a site with not so few traffic, so this is not just a gedankenexperiment, it works. Of course, it is not as solid as captchas, but faster, easier and more user friendly. And it is at least as good as the mathematical problem measure, if not better.

P.S.: Sorry that I still have the math question in my blog comment form. I did not write the Flatpress plugin implementing the empty field yet. But I will.

It is frustrating. Most of the times it would be so easy to fix. What the hell did the developers think about while implementing those “features”?

[Read More…]