Last modified: 2014-09-13 17:11:48 UTC
Short description: The Wikimedia services are a great contribution to the society. Huge database and lots of things to learn, all for free. Still people don’t use Wikimedia services to its full extent and feel lazy about browsing to different places. So it came across my mind to connect all the services to produce best results at one place. I propose a project which has the potential to bring all the services together and produce handsome results from all services (Wikipedia, Wikiquote, Wiktionary , etc. ) dynamically , suggesting user the ways and methods to dig into Wiki Database and search appropriate data and also making Wiki a better learning hub by reflecting the most often searched term or topics on the page so that people can come to learn besides their own subject matter too. SYNOPSIS 1.) Dedicated search page with a perfectly functional search engine which can provide a descent way to search through all Wikimedia services being offered to users till the date. One may name it Wikisearch. 2.) People generally come to Wikipedia while they are searching for any relevant term of use through Google. Changing people’s perception of Wikimedia by developing an exclusive search facility to dig in WIKI database so that they directly approach Wiki Search. 3.) Introducing the best and popular searched queries or terms or topics and reflecting them on search page so that people can come to Wikipedia not only when they want to search a particular topic but also when they want to learn besides their own subject matter. Providing learning and education free of cost is the priority of Wiki and always will be. 4.) Using wiki to its full extent by providing a centralized search engine with capability of searching through every Wikimedia service and accumulating all best possible search results at one place. 5.) Suggesting user the method of writing his/her query in the search input box and making sure that user approach exactly where he/she intended to. Special keywords can be used to fetch the appropriate results. 6.) Saving every searched term and keeping track of them at every time. By keeping track I mean to say, looking for the most often used terms and the current burning issues being searched for and showing them on them on search page. 7.) The idea involves providing user a place where all the latest happenings of Wikimedia services are listed (the latest and most often searched topics) and an exclusive search facility to dig into the WIKIMEDIA database, producing the best accumulated results at one place. DELIVERABLES Tentative Timeline: Week 1 (Starting June 17) Understanding and studying in depth the database design and schema of current WIKI structure. Week 2 Building the basic ER Model and Data Flow Diagrams and structuring the search engine in sync with WIKI database. Creating the required databases and documenting side by side. Week 3 Designing the User Interface of the whole page. Writing the frontend code and publishing for reviews. Week 4 Starting with backend programming and designing at least 1 phase of the idea per week. Most probably starting with search engine (Phase 1) and developing library functions. Week 5 Backend Scripting for 2nd phase and also keep fixing bugs for previous phases (whether it is designing bugs).Phase 2 involves working on most often searched terms or topics and listing them on page. Working on library functions for phase 2. Week 6 Perfecting all the work done till now and developing proper documentation and SRS representation. Testing all the features developed till now and reporting bugs. Also fixing errors and omitted features if there is any. MIDTERM EVALUATIONS Week 7 Designing error handling techniques and methods for the whole page. Handling errors and providing remedial solutions in real time. Week 8 Backend programming for phase 3, i.e. suggestion list for the query being searched and trying to introduce little bit of Artificial Intelligence in the searching algorithm. Working with the search terms and keywords for appropriate query results. Week 9 The phase 3 might incur 2 weeks. So I will be continuing week 8 work in this week too. Error handling will be simultaneously done with every phase for the whole concept. Also I’ll be publishing the code at every achieved milestone for reviews and feedbacks. Week 10 Testing for all the phases will be done and if there’s any reported error, it will be fixed within this week. Week 11 Finalizing the whole concept and creating the full documentation of the project. Giving finishing touch to the design and the functionality. Most probably the whole concept would be ready to be implemented now. Rest of the details are mentioned on the official proposal page. You can visit here http://www.google-melange.com/gsoc/proposal/review/google/gsoc2013/code_sid/1 I am not accustomed user of bugzilla, so I apologize for any neophytic formats and behavior in the report.
Thank you code_sid. Please provide a link to your proposal in a public URL under your userpage at mediawiki.org. Only mentors can access to your proposal submitted in google-melange.com. Background: Siddharta has submitted this proposal to Google Summer of Code and we have asked him to file it here as an enhancement request for community discussion. If the ideas is found to be good for GSoC then we will need to find 1-2 mentors for it.
CCing Ram. Your opinion here is especially welcome.
code_sid: Have you seen and read https://wikitech.wikimedia.org/wiki/Search ? My very personal opinion: I see things listed here that would take years. Out of scope for a GSoC project of three months, plus the timeline underestimates the complexity of the infrastructure, even if code_sid knew Lucene and https://wikitech.wikimedia.org/wiki/Search already by heart.
Andre Klapper: Yes I have gone through the link and current wiki infrastructure. I think you misinterpreted my idea. I mentioned in the 'short description' to bring all the wiki services together as one common search facility. I am not underestimating wiki search structure rather I am deeply impressed by the complexity of it and I know you guys are still working on improving it. My idea says differently. I haven't planned to build a completely new search algorithm for wiki. This will be like coding google's engine in 3 months. I am going to implement the current search infrastructure in a single input box with capability of searching through all services from one place. For Ex: ---------------------------------------------------------------------- One can search by typing in the box the query and selecting from drop down list the wiki service to search for. The results will be shown instantaneously via ajax just below.If the person is not satisfied with the results he/she can change the service from list and again search at that very moment. or We can show results from all services together in different columns and user can thus judge his/her appropriateness of subject matter. ---------------------------------------------------------------------- According to my experience this won't take long if I will be guided thoroughly by the mentors on current wiki searching techniques and implementations. Also I would like to mention, in my perspective , that summing up all my above mentioned Ideas(there are 3 main ideas...) this page could be the most widely used page by the users.
The current search code (Java) already has a mode for "interwiki" searches but it is currently disabled; enabling it may provide much of the proposed functionality. I'll take a look and post more details later in the week.
Just a note to say that code_sid's proposal can be found as updatable wiki page at https://www.mediawiki.org/wiki/User:Code_sid
@Munagala Ramanath : I think there will be need for a dedicated page for this "interwiki" search functionality to get recognized among the users. How will user get to know if there's interwiki functionality prevailing in wiki structure and moreover there will be a requirement for a good user interface for such a search page to enhance the ease of use. Also talking about the other two ideas I discussed in my proposals: 1) The query suggestions for user to reach at appropriate data set. 2) Listing the most frequent topics in trend. These features will provide user a better place to grow awareness about what going in and around the world. What are your opinions? Keep me posted. Thank You!
(In reply to comment #7) > @Munagala Ramanath : I think there will be need for a dedicated page for this > "interwiki" search functionality to get recognized among the users. How will > user get to know if there's interwiki functionality prevailing in wiki > structure and moreover there will be a requirement for a good user interface > for such a search page to enhance the ease of use. You misunderstood, the interwiki search is _not_ available, only the code is. The feature was disabled years ago. Users didn't have much problems finding it, because it was integrated in Special:Search. Of course after it's restored it may be used more widely and in other places (like MediaWiki:Missingarticle ? ).
Bug 44420 was fixed and the code is now live of Italian language Wikimedia projects. The scope of this bug was never clear and what was clear is superseded; closing.
