Category Archives: Browser

Gnome Yelp is much too slow

If you start GNOME help the first time this takes a lot of time. On my system (1.4 Ghz) I am guessing it needs about 30 seconds (and therefore about 6 times slower than Epiphany) – and you do not get any hint if something is happening. I do not know what they are doing that makes it so damn slow, but my suggestion would be to dump Yelp alltogether and use Epiphany by default. With that decision you get a help faster and also have the full featured browser. Maybe give epiphany a start option like epiphany –help-browser to get in a special mode and thats it. The current status is just unbearable. A help must be available very quick – at best instantly. Everything else does not work at all.

Also I would suggest to minimize the offline help and rather link to the web for uptodate help. Maybe as an option allow people to download the online help as a package. Today it does not make much sense to reduce help to just some official offline help. Like in Ubuntu you get the most help that makes sense in wikis rather than the official documentation.

No offenense to the programmers of Yelp. I just dont know why things are that way – and also I am quicker making a google search and getting an answer than waiting till yelp has started.

1 Comment

Filed under Browser, Free Software, Technology, Web

First impressions of Ubuntu 9.04 beta

I have just updated to newest Ubuntu. Here are some points of interest:

  • Thes removed the shutdown option from system menu. You now HAVE to use the stupid applet. IS this a GNOME or an Ubuntu decision. Personally i dont like this applet for shutting down. I always used the system menu. The applet looks so similar to the pidgin icon. Argh how stupid can programmers be? That would be a reason to dump GNOME. You should not remove essential stuff
  • Flash does not work in Firefox and Epiphany
  • I cant installe epiphany-webkit
  • the volume ruler now is horicontal. Is this better? and he did crash
  • The messaging (if you change volume, network is connected, etc.) now has a black background
  • totem does not play videos and crashes
  • monitor settings look much better. need to test it with my beamer.
  • btw. the gnome 2.26 news again sound silly at least in german like:
  • “GNOME’s web browser, Epiphany, gains an exciting new feature of an improved location bar, similar to the Awesome Bar popularized by Firefox 3.0.” – well Epiphany had this address bar BEFORE Firefox – it might have added searching the titles also – but what kind of GNOME marketing is that?

I am sure there is more to say. Why dont I file bug reports. Because I have lots of outstanding bugs that are undecided ot disputed so I know it does not make sense to submit new bugs. Its more likely that somebody reads this review and acts uppon that as that anybody really cares about new Ubuntu bugs from my experience. Thats sad but its true. That does not mean people do not work on bugs in Ubuntu – but they tend to either fix clear bugs or dismiss anything they dont get. Like the Evolution guys who still have crappy spam marking options. I did file a bug in GNOME bugzilla years ago but they still think spam handing is not important. That was the single reason for me to switch to Thunderbird.


Filed under Browser, Free Software, GNOME, Linux, Technology

In need of a major new GNOME panel

I suggest that people start working on an alternative GNOME panel now. I have seen some suggestions on a GNOME wiki page, but I think most directions are very wrong. Like what you see here:

Essentially these are imitations of the fancy Mac panel. But I think that the Max panel does not give us anything cool as well as the things AWN an Kiba dock do. Look at this video: At one point it shows how to play volleyball with the icons. How stupid is that? I mean cool. Or better: I don’t care!

First of all I still do like the text menus, because you can access a lot of applications and settings without going through a lot of folders and sub folders. But I have some major problems with the panel:

  1. You can fix the position of a panel. but when I plugin in my digital projector the panel moves to the other display (on the right). How can this be called a fixed position?
  2. When the size of the panel changes the position of the fixed icons changes too. I have to resortmany icons after I have dettached my projector display. How fixed are thise positions, then?
  3. So it is impossible to configure one monitor display to show exactly the same things on each occasion. This comes from all the dynamic configuration. At least thats true for Ubuntu. Its like you always plug in a new display which you have never attached before and also like it would make any sense that the panel should never be on the main  display but always on the external display.
  4. You can also not configure to have a second panel which is bound to one display

These are only some of my new points. Here is what I desperately need:

  1. A panel which is much less customizable and dynamic. Because everything that can change results in random results or I have configure or reconfigure the panel. From my view the panel never moved to the point of the rest of GNOME. You can do nearly everything with the panel which does not make any sense.
  2. I suggest that new work goes to a new panel which can be a replacement of the old panel. Maybe one can reuse some of the old code but the essentials should be very different.
  3. I think one very important thing is that screen/display configuration and the panel should be one thing.
  4. Have the ability that the screens (1-4 or so) can be linked to specific displays, so lets say if I have two screens one is the major screen of my notebook (screen 1 on the left) – and the other screen has a different screen size (screen 2 on the right) and is configured for my projector display (which is a 16:9)
  5. If I attach a display and configure the contents, the panel, etc. these settings should be saved for this screen and display so that I get these back once I plugin in that display again. The content (desktop icons) of a display could also be available if this display is detached. Then screen 2 should be reconfigured to a single screen mode.
  6. Essentially if you want to give a presentation you will want perfect control of what the presentation screen looks like and what appears there. If you never know what happens a GNOME desktop can not be used for such a purpose. The frustrating thing is that things rather seem to get worse. I really think about switching Linux distribution because the dynamic screen configuration is really awful. I remember Fedora had a “system-config-display” which worked more relliable. I still dont know why this is not used upstream. Maybe some people think that this dynamic thing is actually good. Maybe it would be if it would work – but till then please keep this as an experimental feature in SVN and do not put it on Ubuntu LTS! grrr. sorry I had to go through a lot of troubles and still do because of this thing.
  7. I would dump all current panel applets because most of them are useless. Instead I would suggest to give a panel some functions like displaying time and weather. Or maybe for advanced users allow them to put a content on the panel which they can insert from script output. Like if I put on the hardware sensor monitor applet I get 10 or more icons on my panel and then have to find out which is the right important temperature. Instead of an applet a user should have a setting where he can enable the display of a temperature and hopefully GNOME can show the right one or give the user the opportunity to to enable the right sensor.
  8. Then there should be an area where the panel displays the icons of the most used user applications. Maybe allow the user to say which applications should never appear. But this would give the user a perfect access to the most used apps without forcing him to put them there. Why should he?
  9. As stated before I think it would be most intelligent if the panel itself is the interface to configure display. So when you add a new screen/display you can choose which panel you want (like no panel, copy major display panel, standard clean panel,…) And maybe have the ability to close/remove a screen with a closing the panel like you do it with tabs in browsers.
  10. I also think organizing screens and applications via the panel should be more intelligent. The tabbed window managers (wmii,dwm,…) invented the ability to group applications – so lets say you can configure a graphic screen and gimp, inkscape, blender,… all open on this one – or you have a mail screen where you work with email. Those screen layouts or definitions could be saved, so that you may have a more general notebook screen but if you go to work and attach your notebook to a large LCD display graphical applications will appear there. Today its rather primitive like you have screens 1-4 and have to move an application there manually on each occasion. Also handling screens should be easy like handling tabulators on a browser. Maybe in the future you may even be able to drag and drop a screen to a remote computer and then the other computer can work or see what you are working on or you can share a screen.

So I think most that is discussed so far on GNOME is nothing more that re-engineering of what Apple did and maybe spice it up a little. Only interesting page on the wiki that I saw was that about GroupBasedWindowManagement. I a pessimistic about GNOME or KDE being more creative in the future. Unfortunately the tabbed window managers still have problems with many applications and often still require some manual configuration. I really think maybe soem new project should try to do things better without repeating past mistakes. Like have less dependencies, so that operating systems like OpenBSD will also follow the development.

Leave a comment

Filed under Browser, Free Software, GNOME, Linux, OpenBSD, Programming, Technology

Using WPA on OpenBSD (2008)

The missing WPA support was something many users where missing on OpenBSD. I now tell some practical steps on how to connect to a WPA encrypted wireless network with OpenBSD. Again – and as often OpenBSD makes it simpler than other OSes. Firstly – during install OpenBSD gives the opportunity to configure networks. I had my Thinkpad R52 not connected to any network while installing. As the internal Intel chip did never work with WPA on my Ubuntu I had a Ralink card inserted. OpenBSD also had some problems, claiming it can not find the firmware. I have not even looked what this means – because I first wanted internet and the Ralink looked much more promising. So Ralink … this is the ral driver. The manual can be found on OpenBSD with ‘man ral’ – on the web: RAL(4). The essential example is here:

Configure ral0 to join network “my_net” using WPA-PSK with passphrase
# ifconfig ral0 nwid my_net wpa wpapsk \
$(wpa-psk my_net my_passphrase)

Anyway. I am not here to copy the manual. What they write there is not wrong – but does not give you any network on a new boot. So the first ral card is called “ral0”. The setup did create a file: “/etc/hostname.ral0”. Here is how its contents looked: “dhcp NONE NONE NONE” – Where do you find more? In hostname.if(5). There you will get the information that you can add “options” behind the dhcp. Those options are the same as the command ifconfig gives.

Back to the example above, what does this do? “$(wpa-psk my_net my_passphrase)” executes the command “wpa-psk” – with the options: 1. SSID and 2. the passphrase. It is able to generate a wpa pre-shared key. You can generate one and COPY that. You can then paste that into the ral0 config file. So a like could look like this:

dhcp nwid <your-network> wpa wpapsk <your-key>

Thats about all you need. And now how you connect to your Router? Nothing simpler. Read about netstart(8). This is actually a non-executable script. You can start your ral0 with “sh /etc/netstart ral0”. And this should be sufficient to get it on every time you boot. A successful connected wpa wireless then will look like this:

$ ifconfig ral0

lladdr 00:80:5a:39:f5:e1
groups: wlan egress
media: IEEE802.11 autoselect (OFDM54 mode 11g)
status: active
ieee80211: nwid <your network> chan 9 bssid 00:1c:10:c1:ab:40 120dB wpapsk <not displayed> wpaprotos wpa1,wpa2 wpaakms psk,802.1x wpaciphers tkip,ccmp wpagroupcipher tkip 100dBm
inet6 fe80::280:5aff:fe39:f5e1%ral0 prefixlen 64 scopeid 0x5
inet netmask 0xffffff00 broadcast

Was that complicated? I would say no. The only think I have not tried is how to best set up multiple networks. Something that works more or less on Linux with the NetworkManager. But it does not work good on my Ubuntu box. I really think that OpenBSD has shown how elegant one can do things. This is now without multiple commands – essentially it is just ifconfig. And why shouldn’t ifconfig be able to do more. Why should we need to have iwconfig, iwpriv,… ?

Linux has tried hard to make some things usable – but on the way to accomplish that Linux developers often throw away old tools and constantly reinvent the wheel. Still maybe it is simple to quickly install an Ubuntu box – but for those who know about Unix and can handle the tools you need to remember less and can do some things MUCH easier. OpenBSD has taken its time to get WPA support – much later than Linux. But now its working. Hope this posting helps some posters to get it done.

As this is an important task I state explicitly that the whole text that I wrote is public domain, so you can reuse it wherever you like.


Filed under Browser, Free Software, OpenBSD, Technology

Linux users not allowed to download Chrome browser?

You might say: well its only for Windows – but thats not the problem. Linux should be ok to run a Windows application with WINE – or even can download via Linux and than have a Windows machine access the setup file. But Google seems to either think that Linux has no WINE for running Windows apps or it does not want Linux users to test – because I click download I dont get anything.  This I guess will work via browser recognition.


Filed under Browser, Free Software, Technology, Web

Googles Chrome looks good

I have read the comic about Googles new chrome browser and I must day I am impressed. It targets some points I like to see in browsers:

  1. They call it “omnibar” – others say Opera was the first – personally I know it from Epiphany for ages – and Epi still is ahead in the implementation because it includes search engines also.
  2. Google Chrome also has tab completion for search engines. Personally I think this is smarter than Epiphanys solution – also Chrome fetches the search engines automatically if you visit a site. That may be difficult if there is more than one search box, isnt it?
  3. Starting with last visited pages is not nice. I dont want anybody tht I show my browser to see exactly which pages I visit most often and so give him an insight about my preferences. This is a privacy issue in my view. I think organizing that in some kind of menu or my recalling adresses also in respect of how often they are visited would be smarter.
  4. The ‘incognito window’ allows me that everything I browse in that window not being saved (no history, etc.). Good idea!
  5. Popups banned to the tabs they come from – yeah. nice.
  6. The security model is smarter than what we had before
  7. A crashing plugin not being able to crash the whole browser would be a great accomplishment. As this is one reason i cant use Epiphany nowadays – because Adobes closed source flash player constantly crashes it. You know that flash nowadays really is on nearly every page – so that happens quite often – how I would love open source flash plugins to become a perfect alternative. Then bugs would be fixed finally.
  8. Chrome downloads malware sites. So this is kind like the adblock lists available? If so, thank you and this can be incorporated into other browsers, too

All in all I think this is a nice thing. There are some other important points, too such as a focus on stability and a fast javscript engine – but thats not really my focus. Why does Google do that? I think that Google has hoped that other browsers would become more capable and that this did not happen fast enough and so they decided o take WebKit and make it even better, just like Apple did with Safari. Again a lesson in Free Software. Where this will end up is that these browsers will gain something:

  • Konqueror uses KHTML, WebKit comes from KHTML. I guess it will use webkit as default in the future, also.
  • Safari as it uses Webkit from beginning or better Apple forked KHTML to do WebKit.
  • Epiphany as it will switch to WebKit
  • Some mobile browsers like for Google Android and Nokia stuff (Nokia bought Trolltech who build Qt. And Qt is used to make WebKit)

Mozilla has lost the support of many projects who were using the Gecko engine or where evaluating it. As far as I understand those projects felt that Mozilla cared more about Firefox and Thunderbird and so they had some bugs open in the rendering engine which they then handled in the interfaces rather in the engine itself – and that Gecko was more complex to use.

I have seen that myself – when you try to build Epiphany and have a constant problem with the XULRunner component (compatibility issues).

So I am seldom critical to what Google does, because I always thought what they did had a good quality. I now have tested Yahoo search for a while because it is default for firefox search in address bar – but the results are essentially worse. And I dont know of any other viable alternative.  And this has been going on for years. Google also has made some smart moves in business like Gmail, etc. where they essentiall daid that disk space should not be an issue and invested some time to make the interface smarter than general web based mailers. In the end Google often wins because their results or products are better.

I agree that they are too dominent now and I would love to see more competitors – but those now are just too greedy and too silly. Google can be beaten for sure – they are also humans (at least I think so, LOL) . There is a lot potential in the web – Some years ago, before Google hit the search engin scene there were actually different search engines you could select from. Today i still only use Google, because I like to get the best result of what I intend for my search. if other search engines would try to do that they would become better, also. But essentially most others try harder in selling ads first and funnily thats why Google earns more money than those others. I think that only Open Source can stop Google -and I even think that Google would not see this as competition. Google already has moved into different fields and soon the search technology will not be the heart of their business. It was just the one thing they were the best in the world and what everybody needed and wanted. So I see the point where Google might even open up the code of their search engines – not soon though. I dont have an overview of where they earn the most but think that still the search engine is too essential for them to share much of the knowledge?

Leave a comment

Filed under Browser, Free Software, Technology, Web

Fixing Adblock Plus on Firefox

As Epiphany on Ubuntu constantly crashes with Flash I needed to go back to usage of Firefox. Its so complicated t use, although it partly has added some Epiphany features like when you type it searches in your bookmarks – what Epi providied since ages. I still dont have illigent search like on Epi where I type a word and then could select a search engine to use. Is there a way to get such search in Firefox, also.

I recently had huge problems after the adblock-plus package got updated – some weird installation error and adblock plus got deactivated – I guess this one problem you get if the core plugins (anybody out there who can live without adblock plus?) are not distributed and manage by the core team. I have deleted these files in my firefox directory: “extensions/*, pluginreg.dat”. That allowed me to install the newest Adblock Plus again. Weird but ok. Problem is that no general user may ever come to this solution. Those who cant handle might be better off when the save all bookmarks to desktop or so –  exit FF – then remove .mozilla/firefox directory and recreate a Firefox profile by start it again – the reimport bookmarks.

I dont really need much extensions. Primarily I only use Adblock-plus and Flash plugin. I still dont get what people do with NoScript plugin – ok you can block Javascript – this can protext you from Javascripts from random websites – but then it seems you need to enable Javascript for every site. My guess is that the people who use this daily do not surf a lot of different pages. Personally I think I surf many thousand different sites in a week – often only for seconds. I really do not get to visit the same sites over ond over again as many users seem to do. So this would mean für me that I would need to add  about 3 -4 seconds for enabling Javascript for one site – then multiply that with the thousands of websites and this adds to some hours a week that it would cost me to select exceptions.


Filed under Browser, Free Software, Linux, Technology, Web