Wikipedia talk:Version 0.8

Page contents not supported in other languages.
Source: Wikipedia, the free encyclopedia.

This page is currently being used to report issues with the ZIM file of Wikipedia Version 0.8, or versions of this in Kiwix, Okawix, etc.

This page was previously used to provide WikiProject feedback on the Version 0.8 article and revisionID selection. That feedback has been archived here.

Please report Version 0.8 bugs or other issues below. Walkerma (talk) 15:12, 24 December 2010 (UTC)[reply]

ZIM File Available?

Is the release available for preview? Where is the ZIM file? RickJP (talk) 05:57, 25 December 2010 (UTC)[reply]

ZIM alpha1 release

I have finished a first version of the WP1 0.8 selection ZIM file. You can download it there:

You may also...

  • have a look to the online version there:
  • download an all-in-one (pre-indexed ZIM + Kiwix + Windows installer & DVD launcher + sources and Ubuntu packages) package which can be directly copied on a DVD or USB Key (especially though for Windows users) there:

Please keep in mind, this is an alpha version.

For the Kiwix software bugs and feature request:

For content issues, please update the following list. Thanks Kelson (talk) 09:49, 25 December 2010 (UTC) :[reply]

Report ZIM issues here

Existence of a search word in an article title should get a very high score, and its existence in the lead section should get a higher score than in the rest of the article; for example, search for composer ranks the composer article only at 6th position, and many less relevant articles before actual composers:
  1. john adams (composer) - 100%
  2. john barry (composer) - 99%
  3. timeline of musical events - 99%
  4. list of major opera composers - 98%
  5. ambient music - 97%
  6. composer - 97%
  7. pere lachaise cemetery - 97%
RickJP (talk) 08:53, 26 December 2010 (UTC)[reply]
The answer to this request is not easy. The ranking system is an heuristic which is like we can see not perfect. They are different weights for title, keywords and content terms. But there is no difference beetween terms at the beginning of the content and terms at the end. I'm sure we could be better, but I think the results are currently pretty good and I have other priorities. Someone wants to work on the subject and modify the Kiwix code ? Kelson (talk) 10:16, 22 February 2011 (UTC)[reply]
  • Musician seems to be erroneously included: rated Stub-class, Low-importance in Wikipedia. RickJP (talk) 16:18, 26 December 2010 (UTC)[reply]
  • IPA chars in template:IPA usually display, however not e.g. in the notes following the table here. They usually display correctly in this table, but not always, e.g. not with "[ˌʔadlo̞ˈjada]". Dan 16:22, 27 December 2010 (UTC)[reply]
  • We do not have at the end of each article a footer with license, link to the original version, etc... This would be better to have one and to refresh the old one of 0.7. Kelson (talk) 16:23, 29 December 2010 (UTC)[reply]


  • Search results are all in lower case (Fix @Kelson). Strictly from a user perspective it may be unimportant, but it will lay us wide open to criticism from academics such as: "The search results are all in lower case, even in cases like Óengus I of the Picts (shown as oengus i of the picts!). Have the text messaging kids taken over Wikipedia?" Can this be fixed? Walkerma (talk) 05:40, 28 December 2010 (UTC)[reply]

  • Downloaded again and it works. RickJP (talk) 13:39, 3 January 2011 (UTC)[reply]
  • There is no welcome page. (Fix @Walkerma) Kelson (talk) 09:49, 25 December 2010 (UTC)[reply]
    • Oops! I'll write something this weekend! Walkerma (talk) 17:18, 25 December 2010 (UTC)[reply]
This page should be OK, I think. Can you check that it looks suitable? Thanks, Walkerma (talk) 05:25, 28 December 2010 (UTC) BELIEVED TO BE FIXED[reply]
FIXED Kelson (talk) 13:06, 13 February 2011 (UTC)[reply]
  • The HTML title tag of the article "Heroes" has a problem. (Fix @Kelson) Kelson (talk) 09:49, 25 December 2010 (UTC)[reply]
    • I see that it doesn't show up in the index under H, and that the title at the top of the screen is funny. Is that the extent of the problem, or is there a more serious issue that I've missed? Can we fix the indexing, at least? Walkerma (talk) 17:18, 25 December 2010 (UTC)[reply]
    • I have reformulated the problem... but I think also that in the index it should be at the letter H. Maybe you should add this issue as a new point. Kelson (talk) 17:54, 25 December 2010 (UTC)[reply]
FIXED Kelson (talk) 13:08, 13 February 2011 (UTC)[reply]
  • They are empty paragraph. Empty paragraph should be remove (Fix @Kelson) Kelson (talk) 09:49, 25 December 2010 (UTC)[reply]
    • I haven't seen any - can you give some examples? Walkerma (talk) 17:18, 25 December 2010 (UTC)[reply]
    • I have reformulated the problem, these are no "articles" but "paragraph". I will fix this issue.
FIXED Kelson (talk) 15:11, 13 February 2011 (UTC)[reply]
  • They are linkless "detailed articles", "see also" (Fix @Kelson) Kelson (talk) 09:49, 25 December 2010 (UTC)[reply]
    • I don't recall seeing that problem in Version 0.7, so what did we do differently this time? Do we need to add some code to remove this sentence/section when the article is not linked? Is that technically feasible? Thanks, Walkerma (talk) 17:18, 25 December 2010 (UTC)[reply]
    • Not sure exactly why they are still there... but I will fix also that point. Kelson (talk) 17:59, 25 December 2010 (UTC)[reply]
FIXED Kelson (talk) 13:12, 13 February 2011 (UTC)[reply]
  • Non-standard characters in the index: Articles beginning with nonstandard characters are missing from the index. Examples include Æthelbald of Mercia and "Heroes". These were all listed correctly in the 0.7 index. Walkerma (talk) 05:40, 28 December 2010 (UTC)[reply]
    • My mistake - they are simply in a different place than in Version 0.7. CBM used an improved system for the index, and now the non-standard characters are assigned a place in the main index. I think the new index assignments are much better than the old one. NOT A PROBLEM Walkerma (talk) 17:14, 29 December 2010 (UTC)[reply]
  • Bad search result for "The Pianist (2002 film)": "<i>the pianist</i> (2002&#32;film)"; clicking on it doesn't bring up any page. RickJP (talk) 13:56, 4 January 2011 (UTC)[reply]
FIXED, this was related to the HTML tags in the title, like for the article "Heroes" Kelson (talk) 21:31, 15 February 2011 (UTC)[reply]

Beta release - please test!

There is now a beta Kiwix release available for Version 0.8, which hopefully fixes all (or nearly all) of the bugs outlined above. Please test this and report back here ASAP, so we can publish this release! Walkerma (talk) 17:02, 16 February 2011 (UTC)[reply]

Online version
ZIM file
ZIM file packaged with Kiwix


Please give us your feedback here. Thanks! Walkerma (talk) 17:02, 16 February 2011 (UTC)[reply]

  • Is there any way to sort the year articles chronologically instead of alphabetically? For example, "95" comes after "949". Also, why is "Computability theory" filed under "R"? —Designate (talk) 17:57, 16 February 2011 (UTC)[reply]
It may be too late to fix the dates for this release, but we'll add it to the bug list for next time. I agree - I hate when indexes do that! I'll look into the R thing - one of our team contributed a lot to that article so they may have some insights. Thanks, Walkerma (talk) 04:46, 18 February 2011 (UTC)[reply]
  • (Fixed in RC1, I was testing with alpha1) The movie search results, such as Toy Story, and perhaps other search results, appear with html italics tags around them. —Preceding unsigned comment added by (talk) 22:52, 16 February 2011 (UTC)[reply]
Can you describe what system & OS you're using? I haven't seen that problem. Thanks, and keep the feedback coming! Walkerma (talk) 04:46, 18 February 2011 (UTC)[reply]
It looks like to be the same issue as for "The Pianist (2002 film)". This issue should be fixed. Have look there . I made an error in, I made the package with the "old" ZIM file. I only fixed the issue yesterday... so I guess you habe the old alpha1 content. Kelson (talk) 05:08, 18 February 2011 (UTC)[reply]
  • Looks good!
  • Has anybody looked into the issue of giving a higher search score to article title and lead section? (see above)
  • Is there a place for small typo fixes? For example:
"1925 – Cal Tjader, American vibraphonist, percussionist and panist (d. 1982)"
"... An unexpectedly long history of sexual selection in birds-of-paradise". Evolutionary Biolgy 9 (235).
(they are fixed in the live on-line wikipedia)
RickJP (talk) 18:50, 18 February 2011 (UTC)[reply]
I think these will be errors in the original online articles. We try to pick the best version of each article, but with about 100 million words in this collection, a few typos are inevitable, unfortunately! Thanks, Walkerma (talk) 21:25, 27 February 2011 (UTC)[reply]
  • Kiwix Issue: The "Search results" sidebar should display the search string. RickJP (talk) 22:13, 18 February 2011 (UTC)[reply]
You are right, but we will remove soon the result sidebar and replace it with a result page "a la Google". Kelson (talk) 10:21, 22 February 2011 (UTC)[reply]
Ah, good! Please let us know when done. RickJP (talk) 19:54, 22 February 2011 (UTC)[reply]

Main page

It should probably link to the

WP:1.0 page somewhere just for context. —Designate (talk) 19:34, 18 February 2011 (UTC)[reply

Beta release of Okawix - please test!

There is now a working version of version 0.8 bundled with Okawix, an offline reader from Linterweb that is an alternative to Kiwix. Please can you take download this and see if it's OK?

This version is for PC (and Mac too, I think). There will also be iPhone/iPad and Android versions available in mid-March. Please leave any feedback below. Many thanks! Walkerma (talk) 21:31, 27 February 2011 (UTC)[reply]

  • In my test, it was very slow to load (perhaps 10 minutes). After that I had only one major problem - at one point, when I tried to do a search, it froze and said "XULrunner is not res

ponding" - then the software crashed. When I restarted, it came up immediately and worked fine. I found out that History only shows a search history, not every page I've been to, but I suppose that's what the browser's back button is for. So, I think it's not perfect, but workable enough for a test release. Walkerma (talk) 03:54, 1 March 2011 (UTC)[reply]

  • As mentioned below, I saw some problems in rendering the non-Latin letters - for example Ante Pavelić appears as Ante PaveliÄ (though interestingly it appears OK in the index). When I tried to do a search for Ante Pavelić, XULrunner crashed. Walkerma (talk) 08:00, 3 March 2011 (UTC)[reply]
  • I searched for "spider", and Spider was result number 9. Bob the WikipediaN (talkcontribs) 17:01, 3 March 2011 (UTC)[reply]
    I propose that articles with a title matching the search term ought to yield the highest percent relevance. Bob the WikipediaN (talkcontribs) 17:08, 3 March 2011 (UTC)[reply]
  • When searching for an article such as "Black widow", the average user won't know the title they're looking for is Latrodectus mactans-- a preview of the first sentence of each article in the search results would be very helpful. Bob the WikipediaN (talkcontribs) 17:08, 3 March 2011 (UTC)[reply]
Are you sure that you using okawix because i don't find any problem with our search engine. :) . Pmartin (talk) 07:23, 8 March 2011 (UTC)[reply]
Whoops, no. That was Kiwix. Bob the WikipediaN (talkcontribs) 07:29, 8 March 2011 (UTC)[reply]

Release on March 1st?

I've been proposing that we make the release official on March 1st, and we will supply Kiwix and Okawix download links on that day. We believe that the current collections (called beta), although not perfect, are suitable for release. If you know of any reason for us to postpone the release, let us know. Walkerma (talk) 21:34, 27 February 2011 (UTC)[reply]

Haven't had a chance to check the new version, yet. Hope to get to it later today or tomorrow. Can we wait? RickJP (talk) 11:42, 1 March 2011 (UTC)[reply]
Yes, I'd like us to get a second opinion, at least! Thanks, Walkerma (talk) 21:15, 1 March 2011 (UTC)[reply]
Mmm... Two downloads resulted in a corrupt archive. Okawix did not report an error, just got stuck in "Please wait while your corpus is being installed". Saw the corruption with winrar. I'll try a third time with another browser (the first two were with Chrome).
A small issue, not a test-release stopper, from the on-line version: do we want to remove dablinks and similar header references to articles not included, such as at the top of Percussion mallet ?
RickJP (talk) 04:05, 2 March 2011 (UTC)[reply]
  • Also install from Okiwix's "Download Wikipedia Version 0.8" failed after about 5 hours, without detailing the reason. Is it possible to split the corpus into smaller parts for download? RickJP (talk) 11:10, 2 March 2011 (UTC)[reply]
Indeed, a download with IE worked!
The corpus looks good generally; I vote for release.
Some issues:
  • A match for a search term in an article title is given a high weight in the results, but not in the lead section. Thus a search for composer gives only: 1) Composer; 2) John Adams (composer) and 3) John Barry (composer). Beyond that, only dates; not, for example, Vangelis, Ichirou Mizuki or Antonio Vivaldi. Similarly, a search for "quicksilver" does not give Mercury in the results.
  • There are still some IPA pronounciation character display problems, for example, in "Washington, D.C."
  • Is the "sister projects" box at the bottom of "Kitesurfing" a mistake? It has a false link to "images and media from Commons" - the words "images and media" look like a link, but clicking goes nowhere.
RickJP (talk) 20:42, 2 March 2011 (UTC)[reply]
This is excellent feedback - thanks! These are all non-trivial issues that we will need to look at; however, for most of them it would delay the release by several weeks and (as you imply) they may not be worth the delay for this release (which is a test release, after all!). I'll go ahead and prepare the stuff postings, for release later on today. MANY thanks, Walkerma (talk) 20:49, 2 March 2011 (UTC)[reply]
Two more things:
  • Reminder: User:Kelson's comment above about a general article footer with license, link to the original version, etc.
  • Noticed that it is not only IPA characters that are not displayed correctly. For example, the Hebrew letters, and the year range dashes in "Aliyah" are badly displayed in Okawix. In the on-line Kiwix version they are OK.
RickJP (talk) 08:43, 3 March 2011 (UTC)[reply]
And two more in Okawix, neither a release-stopper:
  • Search for "dance" shows the article "Dance" twice at the top of the results.
  • Extraneous "i" characters for italic article search results. For example:
    • "i Dance Dance Revolution /i"
    • "i Queen /i (album)"
RickJP (talk) 18:33, 3 March 2011 (UTC)[reply]

Bug report: I downloaded this version and loaded it into Kiwix 0.9 alpha 6 (Ubuntu). It works fine, but the index pages (from the alphabetical index) do not load - I only get a blank page when I try to load them.--

* 22:21, 17 March 2011 (UTC)[reply

Do I understand right, you do not see this page ? What happens if you click on the "home" button (launch "kiwix -jsconsole" to display the error console dialog)? Kelson (talk) 10:56, 18 March 2011 (UTC)[reply]
I see that page, but the links in the alphabetical index (e.g. "A – Aeolic Greek") all take me to a blank page.
With the console, I get errors after starting the application (but no problems from a user perspective), but no errors when clicking on links (but the problem mentioned above). The errors I get on start-up are reproduced below.--
* 23:17, 18 March 2011 (UTC)[reply
Error: jQuery is not defined
Source File: zim:///I/gen.js
Line: 10

Warning: function toggleNavigationBar does not always return a value
Source File: zim:///I/gen.js
Line: 327, Column: 4
Source Code:

Warning: test for equality (==) mistyped as assignment (=)?
Source File: zim:///I/gen.js
Line: 274, Column: 47
Source Code:
            while (element = element.parentNode) { 

Warning: function collapseTable does not always return a value
Source File: zim:///I/gen.js
Line: 227, Column: 4
Source Code:

Warning: assignment to undeclared variable mw
Source File: zim:///I/wikibits.js
Line: 1052

Warning: Error in parsing value for 'filter'.  Declaration dropped.
Source File: zim:///I/MediaWiki~Common.css
Line: 979


I have no answer for this bug. This is the first time I read such report for a ZIM file which is not corrupted (should be the case if you used bittorrent)... especially on Ubuntu I work with every day! Kelson (talk) 11:14, 21 March 2011 (UTC)[reply]
OK. So far I've only seen it with the Wikipedia 0.8 ZIM file, not with any other ZIMs.--
* 21:08, 24 March 2011 (UTC)[reply