Version 0.50

I’ve continued the work to get the application more user friendly. This time not the indexing process but the application itself. I know you will love that new feature:

Return to home feature

In fact, some AJAX style code which gives you a list of the first 25 found articles after you’ve entered the 4th letter.

One more feature: A magnification glass gets you back to the home screen. On a long article simply touch the title bar and then the new icon. Ready to enter a new search phrase. Yes, you can return to the top of a web page by touch the title bar. Here is a picture of my beloved article about “Apple”. Also notice that an annoying rendering bug was fixed (the phonetic text link is now displayed correctly):

Search icon and fixed bugs

So grab yourself a copy. It’s already up at google code. Because the indexer has not change (much) you don’t need to build a new index. Copy and start!

P.S.: Hiding the browsers address bar didn’t works on 1.1.3 so I left that feature out.

P.S.S.: Now when I look at the first screen shot I think to myself: The go button isn’t necessary anymore. I will remove that in the next version.

27 Responses to “Version 0.50”

  1. bwd Says:

    Woohoo, great work! The addition of the AJAX search is a huge improvement and makes searching a lot easier. Thank you, it works great. I’m still having some rendering issues (eg. the english Apple Inc page shows “Template:pp-semi-protect” at the top in red, and also in a few other spots in the doc). I’m also having some Article not Found issues…perhaps I will redo the index and see if that helps. Running 1.1.2 on an iPod Touch and using the en version indexed under Windows.

  2. Tom Says:

    bwd-

    There are a couple of mostly minor rendering issues. I know. But in almost all the time you get the info you want. How much employees has Apple?

    May you tell me which articles are not found?

    One of the worst problems is to display an internal link in blue (present) or red (not present). Sure, I can find that out (the code is in but it’s marked out). But that is slow, every link must be checked.

    So maybe one of your targets is not in the database but you can’t see it. Try the online Wikipedia. If the article is still not found: Mail me: warz@gmx.net

    -Tom

  3. bwd Says:

    Yes very minor rendering issues, just thought I’d let you know. Overall it is working great and it’s amazing what you’ve done. I sent you an email about the articles not found (and also posted in the forum yesterday). Thanks.

  4. John Says:

    oh god, i love you. this is such a great application and it keeps getting closer to perfection. thank you so much, i’m gonna donate.

    i would, however, not remove the “go”-button. what if it’s a three letter word that i’m trying to look up and i don’t have that button?

  5. John Says:

    furthermore i think the magnifying glass could be a little bigger so it’s easier to press it (instead of accidentally touching mobilesafari’s google search button above). maybe that’s just me though.

  6. onesilentlight Says:

    This really looks great. You’ve fixed the periodic table article and the new search features are very useful! Thank you so much for putting the work into making the very best offline wikipedia avaliable for the iPod or iPhone.

  7. Tom Says:

    John-

    sure, If the “Go” button can be removed shorter articles must be display in the list. Something I really would love to see. Display speed at the client end isn’t a problem.

    But I have to tweak the search algorithm, otherwise it takes way to long for short titles. But possible, sure.

    -Tom

  8. Tom Says:

    John-

    the image for the magnifying glass is the in following folder on your device:

    /Application/Wiki2Touch.app/daemon/webcontent/icon_search.gif

    Eventually you have to use right click, “Show Package contents” to look into the “Wiki2Touch.app” package (which is nothing more than a folder). Depends on how do you access your iPhones file system.

    If you have a nicer or larger one fell free to send it and I can put it in.

    -Tom

  9. Dexter Says:

    It looks great, but after I downloaded it, I found out that folder with windows version of indexer is empty ! Can you upload it with indexer for windows ?

  10. unruhschuh Says:

    great work. Haven’t installed it yet but i already love the new search feature! Is there any news on images?

  11. Tom Says:

    Unruhschuh-

    yes, I’ve started to work on image support. Actualy I’m writing a program to download all necessary images from Wikipedia. Only the thumbs of course. Everything goes into one large file which contains a directory of it’s contents.

    No sure when I’m done.

  12. Tom Says:

    BTW, I’ve finally removed the “Go”-Button. The suggestion list is display from the first character on.

  13. Sanford Says:

    Very nice. And the suggestion list display from the first character onwards is really very helpful. The application is much faster too.

    However it seems that the suggestion list does not work when searching in Chinese (even when searching for a page with title longer than 4 characters), maybe it’s related to how the indexing is done?

  14. Dexter Says:

    Tom, could you please upload executable indexer for windows ? There is written “The indexer is available as an executable file (Windows version is included)” at description of 0.46 version, but I can´t see any in the 0.50 version package.

  15. Tom Says:

    Dexter-

    Upps, sorry. It’s up as 0.50b.

    -Tom

  16. Sanford Says:

    More debug info about the unicode search problem:
    URL: GET /ajax/search:%E9%A6%99%E6%B8%AF%E5%9C%B0%E9%90%B5 HTTP/1.1

    Returns nothing

    URL: GET /ajax/search:%E9%A6%99%E6%B8%AF%E5%9C%B0%E9%90%B5a HTTP/1.1

    Returns 3 search result
    香港地鐵Adtranz-CAF電動列車
    香港地鐵Adtranz-CAF電聯車
    香港地鐵Adtranz-CAF

    Since %E9%A6%99%E6%B8%AF%E5%9C%B0%E9%90%B5 is 香港地鐵 the 3 search results (and a lot more) should be displayed even without typing the additional “a” in the query.

  17. Tom Says:

    Sanford-

    thanks for checking and providing the info. It fit into the picture: I found out, that I get now result (in the new version, return a list from the first char) if I use one of the German unicode char (”Ä”). Should return at least “Äpfel” (apples). But it returns nothing. But as soon as I type “Äp” I get a result.

    Looks like that there must be at least one “normal” char. Should be easy to fix.

    BTW: Looks like that you got the new indexer running. What was the problem?

    Thanks!

    -Tom

  18. jim123 Says:

    your instructions are sort of vague. I put the wikipedia app in the /applications and it shows up on my spring board. But then it says :

    4)Put everything in the “iphone-ipodtouch/private/…/Wikipedia” folder to the “/private/…/Wikipedia” folder on your device

    What is the whole directory name of private/…/Wikipedia b/c i dont know where to put the file.

    And where do i put the wiki articles and do i extract them

  19. Sanford Says:

    Thanks for looking into this problem. I think the problem is probably coming from the capitalizing routine you have in your program.

    Since you have provided the Windows based indexer I just did the indexing with a Windows computer… but I will look at the Linux problem later.. I still think you cannot cast _G_fpos_t directly into a long because it is a complex struct on Linux (last I checked it has like 4 members).

    Furthermore, it would be perfect if you can implement the Simplified Chinese Traditional Chinese matching function as well.. The function sounds difficult but is actually very similar to case-insensitive matching you have already done.

    For example typing “a” matches all titles starting with “a” and “A” this is already implemented in some sort of capitalization routine I believe.

    However inputting “Ä” you should probably check it’s lower case equivalent in German as well, correct? Simplified Traditional Chinese matching is exactly like this, the only difference is that you are not matching for upper/lower cases but for 2 different flavours of Chinese (used in different regions) but this is irrelevant to you because the relationship is still 1 to 1.

    e.g. “香港地鐵” should match “香港地铁” where the characters “香港地” in both flavours are the same, but the traditional 鐵 has a simplified version 铁

    This will be very useful as the wikipedia pages are created by Chinese people from different regions and not all of them has redirection pages between traditional simplified

    I can find you a matching table in the next few days.

  20. Tom Says:

    Sanford-

    I’ve found the bug. It was the procedure which decodes url encoding (like %20 or so). If such a char was at the end of the url it was not decoded. It is working now very well and you get the list after the first char.

    Send me the table and I do “lower case” for chinese characters. I’ll try to send a email to the address you’re using here so you can sent me the list.

    Yes sure, there is a method which calls “utf8_to_lower()” which does that. But for the latin special characters (like the “Ä”) you don’t need a table. Simply add 0×20 between 0xc0-0xdf. That does the trick.

    -Tom

  21. Lourenço Says:

    Hey, when i was indexing i got this:
    repackaging enwiki-latest-pages-articles.xml.bz2
    Processed: 18% (584704 articles)
    Repackaging done:
    2791243776 bytes before, 584926 articles written (2497162014 bytes)
    330 articles skipped (130396 bytes)
    2814 blocks
    reading articles titles
    lowering and indexing articles titles
    sorting index
    checking index: passed
    writing index: 2339704 bytes
    I tried it several times and always got the same result, can anyone help me?
    Why is this happening? It always stops at 18%

  22. dionysus623 Says:

    tom thanks for all the hard work. this is by far the best wiki app. but i was wondering if there is a way for me to put multiple article.bins in the same language, i.e. wikipedia, wikiquote, wikibook, etc…

  23. Jim Barr Says:

    Great application! I was originally sold on Wikipedia.app, then saw this and had to try it. The whole concept of being able to have the text content of Wikipedia articles in a pocketable, offline format simply blows me away, and for me, it absolutely justifies the cost I paid for my iPod Touch. While Wikipedia.app got me interested, Wiki2Touch really raises the bar in function.

    What makes Wiki2Touch really stand out is that it uses Safari to display the articles. This means that I can now use Safari’s Landscape mode for easier reading, and I can use Bookmarks, Web Clips, and History to manage searching and reading. This is SO nice!

    I particularly like the fact that the articles are returned as valid URLs, so I can simply use Safari’s Bookmark feature to “save” articles for later reading. I created a Bookmark folder called “Wikipedia articles to read”, so if I come across an article that I want to read, but I don’t have the time to read it, I just Bookmark it into this folder. Later, whenever I have time to read, I have a nice list of articles that interest me to read right in the Bookmark folder. The list jogs my memory, makes finding my articles of interest simple to find, and I don’t have to do a search again just to read them. And I’m done with the article, I can just delete the Bookmark.

    FYI, I’m writing a review of both Wiki2Touch and Wikipedia.app for my new “iPod Touch Tips” site ([url]http://IPTtips.com[/url]) and I hope that my review of Wiki2Touch helps promote it more!

    -Jim Barr
    http://JimsTips.com
    http://IPTtips.com (iPod Touch Tips)

  24. Tom Says:

    Dionysus-

    sure, that is possible. For any database create a “new” two letter language code. Digits should also work but be sure to use only two letter codes.

    I.E. use “wq” for wikiqoute, put the articles.bin into a subfolder “wb” and add the “language.config” from the directory of tjhe proper language. I.E. use the “language.config” from the “de” folder if you’re using the German “wikiquote”.

    Access the articles in the web frontend by prefixing “wq:” in front of the articles titles.

    Please drop me a not if it’s working for you.

    -Tom

  25. Auge Says:

    helo,
    i get an error when i opne wikipedia

    SSH-2.0-OpenSSH_4.6 Protocol mismatch

    When i delete open SSH from the install than wiki wokrs as normal.

    with 0.5 the app opens very slowly why?

    thank you very much

  26. Tom Says:

    Auge-

    I get the same error when I try to open a page on port 22, which is the default port for SSH. So it looks like that you change the port Wiki2Touch is using. The default port 8080 is fine unless you know why you want to change it. 8080 is the second standard port for web applications. The first is 80. I’ve chosen 8080 if someone likes to install Apache. And changing the port to an arbitrary number isn’t recommended.

    Don’t know why the app opens for you very slow. What is very slow for you? Maybe you should consider to reset your device. For me it looks like that you’ve tested a lot and installed/uninstalled a of programs. And there are some program out which can really get your device into trouble.

    -Tom

Leave a Reply