Linuxpedia project
Rationale
There are as many Linux information systems as distributions. Linuxpedia is an effort to interlink those information systems with the objective to improve the Linux user experience and the "Linux process".
Some quotes by Mark Shuttleworth
[...] the ability to say, "OK, there's this bug, and it shows up in Debian, Firefox and Ubuntu." And you can track the status of that in different places, and go in and edit it in different places.
"I can see immediately when a bug is fixed in a distro, and you can say, "Show me all of the bugs which are fixed somewhere else", because that's low-hanging fruit. I can immediately go in there, take the patch and merge it."Source
"To glue our pipeline together, we need tools. Collaboration is an easy term to say, but it's hard to do. We often don't know who to talk to upstream, so the question is: how can we make collaboration better?"
"Many people work on bugs in many distributions, but they're the same bugs, being seen by different pairs of eyes. It's an open-source saying that 'Many eyes make bugs shallow' but we should get all the eyes together. It's the same thing with patches, Shuttleworth continued. "We need to make it easier for developers, across projects, across distributions, to work together."
"At Ubuntu, we use LaunchPad to try to track bugs in a centralized manager, but it doesn't work as well as we would like. We need a federated, decentralized system to track bugs and patches. When a bug is fixed in a federated system, all of the other systems can pick up on the fix," said Shuttleworth."
"Bug tracking, the number one problem is to make it easy for people to turn in bugs. If you have to have a personal relationship to get bugs reported, that's a problem. We also need standards for bug descriptions."
"We can't expect free software developers to agree on a single bug or patch tracking program, but we can create secure open APIs (application programming interfaces) between Bugzillas. These, Shuttleworth called "Porous federated containers."Source
What's the innovation?
- There are attempts to centralize the Linux documentation (TLDP.org, Wikipedia free software portal are good examples). Those sites focus on documentation, not on information system interlinking. They won't let you browse all the bugs related to a given driver in all the main Linux distributions for example, or view all the K3b tips referenced in the various Linux wikis.
- The hyperdata web is emerging. Large scale Semantic Web applications remain uncommon. Linuxpedia will be a playground for putting into practice the semantic and social Web principles. DBPedia and Freebase are among the most well known large projects harnessing the Semantic Web approach. Linuxpedia will be more focused than those two: it will describe only the Linux ecosystem, as accurately as possible, taking advantage of existing vocabularies. Freebase has preliminary support for collective ontology design. Linuxpedia aims at experimenting with collective ontology design as well.
- Linuxpedia will focus on metadata rather than on pure data. The objective is not to centralize the contents related to Linux on a specific Web island. It's more to interlink existing data using metadata.
- Linuxpedia will consist both of a knowledge base and of a social network.
- Linuxpedia will include mashups combining data from external providers of Linux related data such as http://www.linux-usb.org/, http://tuxmobil.org/, http://www.linux-laptop.net/, http://hardware4linux.info, http://tldp.org
- Linuxpedia will use community detection algorithms and will include group forming and social search features
What problems will it solve?
Linuxpedia is meant to help users, developers, software vendors, service providers answer the following questions:- How can I discover new applications that will fulfill my needs? How can I know quickly what the others say about that application? Especially people I trust.
- How can I know more about the bugs related to Eclipse in all the Linux distributions?
- What are the events in a way related to Linux software that will take place in Europe in 2008?
- How can I find easily learning and howto material on anything that relate to Linux?
- Where can I get help about any issue that relate to a Linux distro, whatever my distribution is? I know my problem may happen in various distributions, but I don't want submit my question to all the distro specific forums and mailing-lists.
- How can I know which hardware is running fine with a Linux OS?
- How can I easily enhance the examples available in the man pages?
- How can I get advanced contextual help about a given command directly from the command line?
- How can I get contextual and accurate help from an up to date knowledge base about a given application directly from the application?
- How can I find an expert whom I can trust who will be able to sort out a complex problem urgently?
- What metrics are available to compare several Linux distributions?
- Which companies provide services around Linux? What service specifically? What do the others think about them?
What type of contents will host Linuxpedia?
Linuxpedia should include at least the following contents:- Man pages in various languages, with one wiki page for each man page for storing examples
- Browser of Linux packages, with relevant links to distribution specific pages (for instance the k3b page will also include links to the pages of k3b packages on the Web sites related to Debian, Mandriva, Redhat etc.)
- Wider index of the Linux-related Web than what is currently available on Wikipedia Free software portal (see how many pages are currently available on it)
- Linux related quotes, events, people, companies
- General help desk with links to the Linuxpedia wiki
- Social network between the members
Linuxpedia will contain mashups
Linuxpedia will include mashups based on external data providers such as:- usb data
- http://hardware4linux.info/
- http://www.linux-laptop.net/
- http://hardware4linux.info
- http://qsos.org/ ("QSOS is a method, designed to qualify, select and compare OSS")
- http://tldp.org
- http://del.icio.us/tag/linux
- XWiki Watch for Linux news: collaborative annotation of feeds related to Linux
- Facebook Linuxpedia group
Ontologies
The following ontologies are good candidates for being used by Linuxpedia:- FOAF and SIOC for describing people
- SKOS
- NEPOMUK ontologies for annotations
- DOAP for describing software projects
- hwlister schema for describing hardware
- Baetle for describing bugs
- Collaborative design of complementary ontologies needed to cover the Linux world
Why not use Wikipedia Free Software portal or TLDP.org?
- Linuxpedia will more relate to metadata than to data. The objective is not to centralize the contents related to Linux on a specific Web island. It's more to interlink existing data using metadata. Those links will be created using Semantic Web standards. Linuxpedia will include links to Wikipedia entries, or will include the contents dynamically.
- Wikipedia is not the right place for storing all bash commands as top level pages with links
- we want to put into practice a P2P wiki with semantic capabilities because we believe that the advent of P2P and semantic functionalities in wikis is bringing all sorts of innovations in the way knowledge is created an used. The Wikipedia project is widely experimenting with such functionalities already, but given the complexity of the challenge it's clear that several initiatives are needed. P2P aspects will allow to scale well at low hardware costs; semantic aspects will allow the creation of a Web of data.
- we want to have a social layer on top of the wiki: authors should be able to tag and rate each other
- the wiki will not only gather contents, it will also contain applications and integration with external content sources. Typical applications will include: a social network, customizable calendars related to Linux, a help desk system focusing on Linux issues, a bridge to elearning management systems, a software catalogue with specific features related to installation
- there may be various content licenses
- the contents won't relate exclusively to free software: they will relate to things that relate to the Linux based operating systems. This will include proprietary software running on Linux, marketing events related to Linux etc.
- application mashups: need for an application platform. MediaWiki is not enough.
- playground for experimenting Semantic Web principles.
Why Linuxpedia instead of FLOSSpedia?
- We want also contents about proprietary software running on Linux.
- The Linux community is both wide enough and focused enough for a fruitful exchange of contents and set up of an accurate vocabulary
Infrastructure
Linuxpedia will rely on XWiki:- XWiki is an application platform (lets us create quickly applications)
- XWiki is involved in Nepomuk: semantic features are being added to the engine
- XWiki will have P2P and offline support through
- through the XWiki Concerto project, XWiki will have P2P support by 2008, allowing for low cost hosting maintenance
- the Curriki e-learning experience can be reused
Version 1.2 last modified by Arkub on 18/08/2008 at 17:56
Comments: 0