People like to search – and often people like to search for specific stuff – like a wikipedia article, the Amazon product base, a movie in IMDB or in social bookmarking. I have thought about that for a while. Firefox lets you install search plugins to be able to select this more comfortable and Epiphany also allows you to define “intelligent bookmarks”. But is that all really intelligent?
- You should not need to install a random plugin on your system or browser. An installation is like an operation of a human – there is always a chance that something goes wrong or you get infected. Also what Firefox gives you is a selection of websites. The options to search are: ENDLESS. Which means that the search plugins menu could be endless and you could have a never infinite list of plugins.
- Defining intelligent bookmarks isnt always easy, especially when its not simply URL based but hidden in a search form.
How would a really intelligent search work?
I don’t know how you search but I often do something like this: I look for a technology or a product – lets say I search for an USB microphone – this means I need to know what makes up a good microphone – I need some customer opinions – and also a comparison of prices. In this search might be involved: tech sites, Wikipedia, review sites(like dooyoo or ciao.com), online shops,… . The problem with search with a general search engine is, that it doesnt understand my search. How could it? I think thats only possible with user collaboration and when the users give feedback about their search. The problem with that is that people are leaving a search site when they browse other content. What could we do? I think the only possibility is to integrate intelligence into the browser itself. I should be able to save “search paths” into my browser – maybe not bookmark a page but mark a sentense that gives me an answer and link that to the question. So you might start with just typing in a question in you local browser. This now uses a desktop search engine to look up if there is a similar question. So you might even get saved answers by typing in the question – like “How much euro is 1 us dollar” or you even could say “euro dollar”. Typing in “time” could give you the time. or when the desktop would not know what you mean or you explcicitly tell it to search online it could try to identify your search like:
- euro, dollar – both are currencies – so the user most likely wants to see their relation.
- usb microphone – the desktop could know or maybe lookup in some databases that this is some technical product – question would be if the user wants to understand how they work, want to get this working on his computer just buy such thing or get some recommendations.
- About recommendations: Users could interactively say what recommendations they like – or could trust some users (friends, colleagues) what recommendation sites might be helpful.
Maybe Wikia is now on the way of implementing this – but personally I strongly believe that the important part has to be the browser or desktop search engine. And then it can link to specialised searches like Technorati – but maybe rather fetching the content than opening a web site. I think opening a website should be the last thing to do. I dont think it makes much sense to load tons of websites on a local computer without any need of all that material – and also – why loading a web page, stripping out the adds as good as possible and then search for the real content. This is all because of too much crappy business models based on advertisements – while this all takes much more of our valuable time and makes getting the information we want or need much too hard.
Ok Adam wanted to have some feedback about his latest post about Apples and Oranges. I will take on this task. Lets start with: I could not disagree more:
- Adam thinks that Microsoft is successful because its monolithic
and that Linux should be, also. I don’t think that would be the right way. First of all Linux is so successful because it was not monolithic but was able to adapt in different environments. It gave people the freedom they did not get from Apple or Microsoft.
- Slow releases I think is the most problematic thing we still have. Release often and early is better for an increased development velocity. To be honest: Linux is still very much in the flow – that means changing APIs and backwards incompatibility. Is this bad? Yes. But this does not have to be this way forever. One way that often turned me away from distros like SuSE or Fedora was their lengthy update cycle – forget about Debian in this relation. This only results in totally boring distributions. The problem is that on Linux you get software via a distribution and on windows you download via web or buy in a store. So the only way in Linux to get updates is via distribution. Of this would be only every 1 1/2 years or more like Adam suggests Linux would always be more behind – more than it already is, often, due to distribution that are, due to their release management systems, not be able to release often (like RPM or DEB based systems). I think the way to go is rather to allow people to just install from every source and also let the users decide if they want this software – and not having only a few central repositorities
- Linux IS a platform? Hell yeah! Linux is available on a wider varity of architectures – more than any other OS. Adam seems to suggest that packages should have less dependency so every package should include all the libraries. That might make sense in some way if one looks at software a bit like appliances so to make a programs less dependent on the underlying infrastructure. This can be done today already. But we should also not forget that the FLOSS way is to cooperate. So if you take GNOME or KDE as a desktop environment they are set up in a way that let applications work together. and it makes no sense to put in redundant packages and dependencies. I think maybe its good to include more small libraries in an application statically because depending on it from the distribution would mean that it would have to be packaged already.
- Novell/SuSE/integration. i could hardly agree that this is agood example because I had a hard time to get everything to work there. And also they misuse GNOME and change things form upstream that should not be changed. So that results in a menu thats really not very usable and instead looks more like Windows XP (Is till do not believe that this really is the result of hard usability research!) and things like the font beeing too small for the panel clock (since many months now and the dont fix it). Foresight to me seems to be much more integrated in the sense that it uses the default web browser of GNOME: Epiphany which should be default on ever GNOME desktop because it is much more integrated.
- Hardware: Here I would partly agree: We should have a Linux hardware standard. Or better free and open hardware standards. Here are some initiatives on the way from manufacturers of printers and wireless nics in collaboration with Linux developers. We could even use some more – not have just Linux compatible or certified hardware but hardware that is made for Linux. This would mean to have a more long term plan so that we coudl tell hardware vendors what standards we like to see in two or three years.
- backwards comaptibility is not important really. Microsoft and Apple never had that, really. Its nice to have ok.
- What i think Linux needs is more of a vision. And that you can depend on some things. So it would be nice to have more stable APIs and a foreseeable future. So that users and developers know where they are heading. Right now this seems to be impossible for the big distros like SuSE,Fedora, Ubuntu. The only distro that seems to be able to release in time wioth GNOMEs 6 month cycle is Foresight and so its the only distro I ca depend on – although it might have some issues here and there – but this I had on all distros with the only difference that they rarely go fixed – and that fixing with a package of my own was never a way to go.
Why did we do this? Why didn’t we continue to present Firefox as the front row browser? Well its simple: We are THE GNOME distribution. We want to show the latest and greatest in GNOME software – and to experience what a real GNOME desktop looks and feels one should really should recommend the official GNOME browser.
- Download http://www.schooner.com/~loverso/no-ads/no-ads.zip
- Unzip the contents (no-ads.pac) in your home directory
- open gnome-network-preferences
- choose Automatic Proxy Configuration
- type in: file:///home/homedir/no-ads.pac (where “homedir”should be replaced by your home directory)
Maybe distributors should think about including such a thing as a standard? So maybe as a shared file for many different browsers. GNOME people always told be nobody wants different proxies for different browsers – so consequently sure everybody wants an adblocker.
I could also think of a preference application that helps to edit such a file. This could also be inside the gnome-network-preferences. I think I am going to file a bug at bugzilla.gnome.org, if it does not yet exist.
I hate things to be only available in CVS, like the Adblock extension for Epiphany web browser. Some developers either think everybody likes to download fom CVS or think their software is not worth publishing (yet). I understand that if people just start, but I think it is strange if things keep being so for years. The one major reason why people would not want to use epiphany is that there is no ads protection. So for many years now people rather use Firefox or Mozilla, because Epiphany does not protect them. Even if the adblock extension would be inferior I think one should at least publish an alpha or beta version so that people can test it. I gues it will not totally break epiphany?
I guess that the developers are often just too prowd to release something inferior – they rather live with thousands of users that switch to Firefox and possibly never come back as that they would present something that does at least do the job imperfectly. I would also like to call it “The GNOME disease” because such things happen in GNOME quite often. Othere people also came up with this terms but defined them differently.
I would describe our “disease” as following:
- The end user experience is very different from that of the developers that always live in CVS. So they are living in a totally different world. While most users are eagerly waiting on some bugs to be fixed, developers live in tomorrow now and always actually use features users just have heard of.
- Playing with usability: Some annoying usablity issues do not get fixed for years because it has no priority in an overall strategy – it is not exciting – so nobody likes to fix it.
- Not-implementionism: Developers take feature wishes as prayers of the community. A good developer god listens but does not act!?
- Meritocracy: The important things are discussed in small circles and not in public. By that one can successfully reduce the people that can get involved and the amount of ideas one has to deal with.
- KDE-hatism: No GNOME dev likes to talk about the other Desktop
No way out? I dont want to be misunderstood: I like some direction in GNOME, but usability also means to listen to the users and ACT, not only listen and think that users are stupid. Partly the usability of GNOME ist stubborn and far away of the users interests. Usability is fine, but only if it is made for REAL users. I am also pessimistic about collaboration with KDE (via Freedesktop).
I tend to hope that someone comes up with a new common desktop that builds on what KDE and GNOME have accomplished but binds more users from both sites. So maybe the new desktop could be FreeDesktop.org? Why am I so pessimistic? Because time goes by and although some interesting things go on inside the desktops we still have not gained much more market share and still developers live in their own worlds.
The task to make a popular and attractive free desktop is not trivial. I fear that ressources are not well and that there are too much different priorities – and none of them really has the user in its focus – even in projects which work a lot on usability. You can kill a community with usability overload! Still the essence is to keep users interested in the applications. So it is important to build an keep some excitement about new features or new ways to do things. And sure this never will happen under GNOME or KDE only in a sense that could eventually attract more people.
Maybe not all of this is true. But I think at least some truth is in all of these points.
Another thing that is happening is that there is a possible swift in the community from pure developer communities to more mixed communitities. My observation is that developers have often huge reservation against newcomers/outsiders. This is also partly because of their love for meritocracy and their scepticism about anarchy. The challenge of todays open source development is to cross borders, to unite communities and to set new standards. Right now many projects like to keept their borders closed or to have them controlled. Every community has their culture which they like to stick to. But that makes participation complicate. Often the outsiders are the openers that can move something if they have gotten excepted by the developers.
Maybe this all sounds a bit like a sandbox play. It is!