Hi,
I know most of you hate reinventing a wheel so I first send it here, before I launch that project :)
Some of you probably know kiwix - kiwix.org which is offline wikipedia reader. I think the idea of this reader is cool, most of you probably sometimes wanted to access wikipedia while being offline somewhere, but couldn't. Kiwix can help with this, however it has one big problem and solution for it is so complex that it would basically need a rewrite of whole thing.
That problem is that you need to download pretty huge file (40+GB) in order to use it for en wikipedia for example. And if you wanted to update those few wikipages you are interested in, to a latest revision, then you again need to download that huge file.
That suck. Especially with GPRS internet and similar connectivity and it also suck because mobile phones don't even have space for so much data. My idea is to create app similar to kiwix, that would use SQLite DB and using wikipedia API it would (slowly, apache friendly) download contents of any mediawiki installation based on user selection, so that you could download just a 1 page for offline reading, or 1 category. Or 1000 categories. Or precompiled sets of pages created by users (books). You could easily update these using API anytime to latest version. You could get media files for these pages, etc, etc... (You could probably even edit the pages offline, and then update them when you are online, but that is just extra feature)
I think this approach would work much better and it's sad kiwix already doesn't support it. At some point, if it worked I think this new code could be merged back into kiwix, I am going to use C++ in the end, which kiwix uses as well.
What do you think about it, is it worth of working on? Is there actually a community of "offline wikipedia readers" that would appreciate it?
Thanks
You don't need a new application for this, you just need some upgrades to the ZIM generation pipeline. How about sending patches for https://sourceforge.net/p/kiwix/other/ci/master/tree/mwoffliner to generate category-based ZIM files and the like?
Customisation possibilities used to be better here before https://www.mediawiki.org/wiki/Offline_content_generator was deployed on Wikimedia projects, as ZIM files of few hundreds pages were very easy to make on wiki. Help with the OCG component is also appreciated.
Finally, icnremental ZIM updates are an area of ongoing work. You can probably help here as well: https://phabricator.wikimedia.org/T49406#492033
Nemo
Ok, see my responses bellow. I am interested if there are some users who actually do use offline wikipedia and how would they like this new approach.
On Fri, Jan 23, 2015 at 1:34 PM, Federico Leva (Nemo) nemowiki@gmail.com wrote:
You don't need a new application for this, you just need some upgrades to the ZIM generation pipeline. How about sending patches for https://sourceforge.net/p/kiwix/other/ci/master/tree/mwoffliner to generate category-based ZIM files and the like?
Which is far harder than creating a new application that wouldn't use this ZIM format. So I don't think so.
Customisation possibilities used to be better here before https://www.mediawiki.org/wiki/Offline_content_generator was deployed on Wikimedia projects, as ZIM files of few hundreds pages were very easy to make on wiki. Help with the OCG component is also appreciated.
Finally, icnremental ZIM updates are an area of ongoing work. You can probably help here as well: https://phabricator.wikimedia.org/T49406#492033
Ok these are all "planned to be implemented" solutions that do not work now, and maybe will never work. I don't really see any option to use either of these.
ZIM format is completely incompatible with what I have proposed. It's one time-baked file that is supposed to be read only, not heavily modified. SQLite format is far better for this design and already exist and works.
My goal is to create application that would be naturally extremely small and super easy to use by users who have no understanding of computers and who don't want to download 40gb files in any way. Even if there was a wiki interface that would allow people bake their own ZIM files, I would find it very complex. If you wanted to add 1 page to your collection, wiki would have to generate a new file and you would have to download it again. Using API instead to retrieve data would be more simple.
FWIW I also wrote a web app called "Nell's Wikipedia" which behaves as you propose: https://github.com/cscott/nell-wikipedia
If you wanted to hack on it, it could use a bit of love. --scott On Jan 23, 2015 7:55 AM, "Federico Leva (Nemo)" nemowiki@gmail.com wrote:
Ok these are all "planned to be implemented" solutions that do not work
now
You didn't even read, did you? Let me quote:
I use this code yes and it works
Nemo
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
web-app for offline use? :o
On Fri, Jan 23, 2015 at 5:54 PM, C. Scott Ananian cananian@wikimedia.org wrote:
FWIW I also wrote a web app called "Nell's Wikipedia" which behaves as you propose: https://github.com/cscott/nell-wikipedia
If you wanted to hack on it, it could use a bit of love. --scott On Jan 23, 2015 7:55 AM, "Federico Leva (Nemo)" nemowiki@gmail.com wrote:
Ok these are all "planned to be implemented" solutions that do not work
now
You didn't even read, did you? Let me quote:
I use this code yes and it works
Nemo
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Petr Bena benapetr@gmail.com writes:
web-app for offline use? :o
Maybe they have a small webserver set up in a remote place with only sporadic Internet access? Maybe they're running a Mozilla's Firefox OS on a phone and calling apps for that OS "web apps"?
On 2015-01-24 2:38 PM, Mark A. Hershberger wrote:
Petr Bena benapetr@gmail.com writes:
web-app for offline use? :o
Maybe they have a small webserver set up in a remote place with only sporadic Internet access? Maybe they're running a Mozilla's Firefox OS on a phone and calling apps for that OS "web apps"?
;) Or maybe they're just using HTML5's offline capabilities?
~Daniel Friesen (Dantman, Nadir-Seen-Fire) [http://danielfriesen.name/]
Yes, this. There is nothing contradictory with "web app" and "offline". It works fine in chrome and other mobile browsers, although we deployed it with mobile Firefox for Android. --scott On Jan 24, 2015 3:00 PM, "Daniel Friesen" daniel@nadir-seen-fire.com wrote:
On 2015-01-24 2:38 PM, Mark A. Hershberger wrote:
Petr Bena benapetr@gmail.com writes:
web-app for offline use? :o
Maybe they have a small webserver set up in a remote place with only sporadic Internet access? Maybe they're running a Mozilla's Firefox OS on a phone and calling apps for that OS "web apps"?
;) Or maybe they're just using HTML5's offline capabilities?
~Daniel Friesen (Dantman, Nadir-Seen-Fire) [http://danielfriesen.name/]
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
There's also XOWA, and the dev answered my question about the pros and cons of XOWA vs Kiwix, at https://www.reddit.com/r/ https://www.reddit.com/r/wikipedia/comments/2f7lzi/your_own_wikipedia/ck6oa07 wikipedia https://www.reddit.com/r/wikipedia/comments/2f7lzi/your_own_wikipedia/ck6oa07 /comments/2f7lzi/ https://www.reddit.com/r/wikipedia/comments/2f7lzi/your_own_wikipedia/ck6oa07 your_own_wikipedia https://www.reddit.com/r/wikipedia/comments/2f7lzi/your_own_wikipedia/ck6oa07 /ck6oa07 https://www.reddit.com/r/wikipedia/comments/2f7lzi/your_own_wikipedia/ck6oa07 which you might find useful.
(Great reference to Nell! I've told 3 people to read http://www.goodreads.com/book/show/827.The_Diamond_Age in the last few days. :-) On Jan 23, 2015 8:55 AM, "C. Scott Ananian" cananian@wikimedia.org wrote:
FWIW I also wrote a web app called "Nell's Wikipedia" which behaves as you propose: https://github.com/cscott/nell-wikipedia
If you wanted to hack on it, it could use a bit of love. --scott On Jan 23, 2015 7:55 AM, "Federico Leva (Nemo)" nemowiki@gmail.com wrote:
Ok these are all "planned to be implemented" solutions that do not work
now
You didn't even read, did you? Let me quote:
I use this code yes and it works
Nemo
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Petr Bena benapetr@gmail.com writes:
That suck. Especially with GPRS internet and similar connectivity and it also suck because mobile phones don't even have space for so much data. My idea is to create app similar to kiwix, that would use SQLite DB and using wikipedia API it would (slowly, apache friendly) download contents of any mediawiki installation based on user selection, so that you could download just a 1 page for offline reading, or 1 category. Or 1000 categories. Or precompiled sets of pages created by users (books). You could easily update these using API anytime to latest version. You could get media files for these pages, etc, etc... (You could probably even edit the pages offline, and then update them when you are online, but that is just extra feature)
Also see the application that WikiEM (Emergency Medicine) is working on. Dan Ostermayer (CC'd -- Link to original[1]) recently applied for a PEG Grant[2] to continue development on their non-Kiwix alternative for offline Wiki reading[3].
They didn't get a grant, but, from past conversations with Mr. Ostermayer, I understand that content from WikiEM is already being used offline in the Emergency Room and he will continue work on this by getting funding elsewhere.
If nothing else, this would probably be a good place to start.
Mark.
Footnotes: [1] http://article.gmane.org/gmane.science.linguistics.wikipedia.technical/81173
[2] https://meta.wikimedia.org/wiki/Grants:PEG/Offline_MediaWiki_search_for_NASA...
[3] https://github.com/ostermayer/offlinexml
Petr, do you think it would be an option to use git version control as a storage format instead of openzim? Which would facilitate edit and merge back changes?
Rupert On Jan 23, 2015 11:59 AM, "Petr Bena" benapetr@gmail.com wrote:
Hi,
I know most of you hate reinventing a wheel so I first send it here, before I launch that project :)
Some of you probably know kiwix - kiwix.org which is offline wikipedia reader. I think the idea of this reader is cool, most of you probably sometimes wanted to access wikipedia while being offline somewhere, but couldn't. Kiwix can help with this, however it has one big problem and solution for it is so complex that it would basically need a rewrite of whole thing.
That problem is that you need to download pretty huge file (40+GB) in order to use it for en wikipedia for example. And if you wanted to update those few wikipages you are interested in, to a latest revision, then you again need to download that huge file.
That suck. Especially with GPRS internet and similar connectivity and it also suck because mobile phones don't even have space for so much data. My idea is to create app similar to kiwix, that would use SQLite DB and using wikipedia API it would (slowly, apache friendly) download contents of any mediawiki installation based on user selection, so that you could download just a 1 page for offline reading, or 1 category. Or 1000 categories. Or precompiled sets of pages created by users (books). You could easily update these using API anytime to latest version. You could get media files for these pages, etc, etc... (You could probably even edit the pages offline, and then update them when you are online, but that is just extra feature)
I think this approach would work much better and it's sad kiwix already doesn't support it. At some point, if it worked I think this new code could be merged back into kiwix, I am going to use C++ in the end, which kiwix uses as well.
What do you think about it, is it worth of working on? Is there actually a community of "offline wikipedia readers" that would appreciate it?
Thanks
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
I don't really know, it is technically possible but probably not suitable. I don't want to create offline wiki. Just a reader of a wiki, so no complex versioning is required for that.
On Sun, Jan 25, 2015 at 6:00 PM, rupert THURNER rupert.thurner@gmail.com wrote:
Petr, do you think it would be an option to use git version control as a storage format instead of openzim? Which would facilitate edit and merge back changes?
Rupert On Jan 23, 2015 11:59 AM, "Petr Bena" benapetr@gmail.com wrote:
Hi,
I know most of you hate reinventing a wheel so I first send it here, before I launch that project :)
Some of you probably know kiwix - kiwix.org which is offline wikipedia reader. I think the idea of this reader is cool, most of you probably sometimes wanted to access wikipedia while being offline somewhere, but couldn't. Kiwix can help with this, however it has one big problem and solution for it is so complex that it would basically need a rewrite of whole thing.
That problem is that you need to download pretty huge file (40+GB) in order to use it for en wikipedia for example. And if you wanted to update those few wikipages you are interested in, to a latest revision, then you again need to download that huge file.
That suck. Especially with GPRS internet and similar connectivity and it also suck because mobile phones don't even have space for so much data. My idea is to create app similar to kiwix, that would use SQLite DB and using wikipedia API it would (slowly, apache friendly) download contents of any mediawiki installation based on user selection, so that you could download just a 1 page for offline reading, or 1 category. Or 1000 categories. Or precompiled sets of pages created by users (books). You could easily update these using API anytime to latest version. You could get media files for these pages, etc, etc... (You could probably even edit the pages offline, and then update them when you are online, but that is just extra feature)
I think this approach would work much better and it's sad kiwix already doesn't support it. At some point, if it worked I think this new code could be merged back into kiwix, I am going to use C++ in the end, which kiwix uses as well.
What do you think about it, is it worth of working on? Is there actually a community of "offline wikipedia readers" that would appreciate it?
Thanks
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
The storage format is very efficient and there is a c library for it : https://libgit2.github.com It should be not necessary to create complex versioning around it.
You plan to store html or wikitext?
Rupert On Jan 25, 2015 6:37 PM, "Petr Bena" benapetr@gmail.com wrote:
I don't really know, it is technically possible but probably not suitable. I don't want to create offline wiki. Just a reader of a wiki, so no complex versioning is required for that.
On Sun, Jan 25, 2015 at 6:00 PM, rupert THURNER rupert.thurner@gmail.com wrote:
Petr, do you think it would be an option to use git version control as
a
storage format instead of openzim? Which would facilitate edit and merge back changes?
Rupert On Jan 23, 2015 11:59 AM, "Petr Bena" benapetr@gmail.com wrote:
Hi,
I know most of you hate reinventing a wheel so I first send it here, before I launch that project :)
Some of you probably know kiwix - kiwix.org which is offline wikipedia reader. I think the idea of this reader is cool, most of you probably sometimes wanted to access wikipedia while being offline somewhere, but couldn't. Kiwix can help with this, however it has one big problem and solution for it is so complex that it would basically need a rewrite of whole thing.
That problem is that you need to download pretty huge file (40+GB) in order to use it for en wikipedia for example. And if you wanted to update those few wikipages you are interested in, to a latest revision, then you again need to download that huge file.
That suck. Especially with GPRS internet and similar connectivity and it also suck because mobile phones don't even have space for so much data. My idea is to create app similar to kiwix, that would use SQLite DB and using wikipedia API it would (slowly, apache friendly) download contents of any mediawiki installation based on user selection, so that you could download just a 1 page for offline reading, or 1 category. Or 1000 categories. Or precompiled sets of pages created by users (books). You could easily update these using API anytime to latest version. You could get media files for these pages, etc, etc... (You could probably even edit the pages offline, and then update them when you are online, but that is just extra feature)
I think this approach would work much better and it's sad kiwix already doesn't support it. At some point, if it worked I think this new code could be merged back into kiwix, I am going to use C++ in the end, which kiwix uses as well.
What do you think about it, is it worth of working on? Is there actually a community of "offline wikipedia readers" that would appreciate it?
Thanks
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Either only wikitext or both, at least until I would get some wikitext to html convertor lib
On Sun, Jan 25, 2015 at 7:41 PM, rupert THURNER rupert.thurner@gmail.com wrote:
The storage format is very efficient and there is a c library for it : https://libgit2.github.com It should be not necessary to create complex versioning around it.
You plan to store html or wikitext?
Rupert On Jan 25, 2015 6:37 PM, "Petr Bena" benapetr@gmail.com wrote:
I don't really know, it is technically possible but probably not suitable. I don't want to create offline wiki. Just a reader of a wiki, so no complex versioning is required for that.
On Sun, Jan 25, 2015 at 6:00 PM, rupert THURNER rupert.thurner@gmail.com wrote:
Petr, do you think it would be an option to use git version control as
a
storage format instead of openzim? Which would facilitate edit and merge back changes?
Rupert On Jan 23, 2015 11:59 AM, "Petr Bena" benapetr@gmail.com wrote:
Hi,
I know most of you hate reinventing a wheel so I first send it here, before I launch that project :)
Some of you probably know kiwix - kiwix.org which is offline wikipedia reader. I think the idea of this reader is cool, most of you probably sometimes wanted to access wikipedia while being offline somewhere, but couldn't. Kiwix can help with this, however it has one big problem and solution for it is so complex that it would basically need a rewrite of whole thing.
That problem is that you need to download pretty huge file (40+GB) in order to use it for en wikipedia for example. And if you wanted to update those few wikipages you are interested in, to a latest revision, then you again need to download that huge file.
That suck. Especially with GPRS internet and similar connectivity and it also suck because mobile phones don't even have space for so much data. My idea is to create app similar to kiwix, that would use SQLite DB and using wikipedia API it would (slowly, apache friendly) download contents of any mediawiki installation based on user selection, so that you could download just a 1 page for offline reading, or 1 category. Or 1000 categories. Or precompiled sets of pages created by users (books). You could easily update these using API anytime to latest version. You could get media files for these pages, etc, etc... (You could probably even edit the pages offline, and then update them when you are online, but that is just extra feature)
I think this approach would work much better and it's sad kiwix already doesn't support it. At some point, if it worked I think this new code could be merged back into kiwix, I am going to use C++ in the end, which kiwix uses as well.
What do you think about it, is it worth of working on? Is there actually a community of "offline wikipedia readers" that would appreciate it?
Thanks
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Dear Petr
On 23.01.2015 11:59, Petr Bena wrote:
Some of you probably know kiwix - kiwix.org which is offline wikipedia reader. I think the idea of this reader is cool, most of you probably sometimes wanted to access wikipedia while being offline somewhere, but couldn't. Kiwix can help with this, however it has one big problem and solution for it is so complex that it would basically need a rewrite of whole thing.
That problem is that you need to download pretty huge file (40+GB) in order to use it for en wikipedia for example.
We provide smaller ZIM files providing selection of articles (have a look on http://download.kiwix.org/zim/). AFAIK, for the same amount of information, we provide the best compression ratio (and consequently potential update speed).
And if you wanted to update those few wikipages you are interested in, to a latest revision, then you again need to download that huge file.
That's a true, openZIM was thought, implemented and optimized for read-only. But the sentence "solution for it is so complex that it would basically need a rewrite of whole thing" is wrong.
Without repeating what Federico already said (later in this thread but earlier last week), we have almost a solution (implemented by a GSOC student) to make incremental upgrades. This solution works well but is a little bit CPU intensive (due to the need to recompress clusters).
A "lazy" version of the patch functionality could fix that quickly (and for small amounts of articles this would not make a big difference). An "intelligent" version of diff/patch might avoid both the cluster re-compression step and the continuous ZIM file growing might also be implemented.
All of this might be implement in a few weeks by a skilled C++ developer without the "rewrite of whole thing", just adding new functionnalities.
I have been working on this kind of stuff since almost ten years and have seen a lot of different offline softwares proposed and part of them published. None of the ones which brought a real added value to the end-user was developed quicker than in a few weeks.
Regards Emmanuel
That is still IMHO highly unusable, let me explain why:
Let's define USER as a computer illiterate that has absolutely no knowledge of how computers work. They got their PC (or mobile) and want to run offline reader of either wikipedia or their favorite wiki, which can be non-wmf thing.
Problems:
1) There may be no ZIM's for the wiki they want to use and they have no idea how to create one. They won't be able to use kiwix here. 2) There may be some ZIM's but these don't contain pages that user is interested in. 3) There may be some ZIM's but these contain too many pages that user doesn't need and thus are too large.
I am trying to tell you that ZIM is a very limited solution, I don't want to say that kiwix is unusable for everyone, it's a very useful solution for many use cases (for example some school computers / libraries with no internet access could have kiwix with full encyclopedia available to everyone), but for USER's (eg. individuals with limited internet access or no computing knowledge) is pretty much unusable. I myself am having troubles getting kiwix even to try it out. I have SSD on my work PC with less than 2gb of free space and at home, with large HDD I have so slow internet that I would be downloading that ZIM file few weeks. And I don't even live in some underdeveloped country, try to imagine how hard this must be for people with really slow internet (< 20kbs etc).
IMHO best solution for this use-case would be to maintain local DB that would contain only data for pages selected by user, which would be downloaded using mediawiki API's, so that there would be no need for any ZIM packaging whatsoever. The local DB might take more space than ZIM, it's possible that compression wouldn't be so effective. But given that user would only have those pages they are interested in, it would be likely much smaller than whole collection. They would make "their own" collection, easily with few clicks. For any wiki. Having this option in Kiwix would be cool, but I can't see anything else what could be reused than the reader itself (the part of kiwix that turns wikitext into html page and display it on screen).
On Mon, Jan 26, 2015 at 12:30 PM, Emmanuel Engelhart kelson@kiwix.org wrote:
Dear Petr
On 23.01.2015 11:59, Petr Bena wrote:
Some of you probably know kiwix - kiwix.org which is offline wikipedia reader. I think the idea of this reader is cool, most of you probably sometimes wanted to access wikipedia while being offline somewhere, but couldn't. Kiwix can help with this, however it has one big problem and solution for it is so complex that it would basically need a rewrite of whole thing.
That problem is that you need to download pretty huge file (40+GB) in order to use it for en wikipedia for example.
We provide smaller ZIM files providing selection of articles (have a look on http://download.kiwix.org/zim/). AFAIK, for the same amount of information, we provide the best compression ratio (and consequently potential update speed).
And if you wanted to update those few wikipages you are interested in, to a latest revision, then you again need to download that huge file.
That's a true, openZIM was thought, implemented and optimized for read-only. But the sentence "solution for it is so complex that it would basically need a rewrite of whole thing" is wrong.
Without repeating what Federico already said (later in this thread but earlier last week), we have almost a solution (implemented by a GSOC student) to make incremental upgrades. This solution works well but is a little bit CPU intensive (due to the need to recompress clusters).
A "lazy" version of the patch functionality could fix that quickly (and for small amounts of articles this would not make a big difference). An "intelligent" version of diff/patch might avoid both the cluster re-compression step and the continuous ZIM file growing might also be implemented.
All of this might be implement in a few weeks by a skilled C++ developer without the "rewrite of whole thing", just adding new functionnalities.
I have been working on this kind of stuff since almost ten years and have seen a lot of different offline softwares proposed and part of them published. None of the ones which brought a real added value to the end-user was developed quicker than in a few weeks.
Regards Emmanuel
-- Kiwix - Wikipedia Offline & more
- Web: http://www.kiwix.org
- Twitter: https://twitter.com/KiwixOffline
- more: http://www.kiwix.org/wiki/Communication
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
On 26.01.2015 13:09, Petr Bena wrote:
- There may be no ZIM's for the wiki they want to use and they have
no idea how to create one. They won't be able to use kiwix here.
We propose a ZIM file for most of the "important" projects. The problem is that we still don't have the resources to generate all of them one time a month. But, this problem is going to be fixed, we are currently building a farm in wmflabs to provide one time a month a fresh ZIM file of all projects.
- There may be some ZIM's but these don't contain pages that user is
interested in.
- There may be some ZIM's but these contain too many pages that user
doesn't need and thus are too large.
I am trying to tell you that ZIM is a very limited solution, I don't want to say that kiwix is unusable for everyone, it's a very useful solution for many use cases (for example some school computers / libraries with no internet access could have kiwix with full encyclopedia available to everyone), but for USER's (eg. individuals with limited internet access or no computing knowledge) is pretty much unusable. I myself am having troubles getting kiwix even to try it out. I have SSD on my work PC with less than 2gb of free space and at home, with large HDD I have so slow internet that I would be downloading that ZIM file few weeks. And I don't even live in some underdeveloped country, try to imagine how hard this must be for people with really slow internet (< 20kbs etc).
What you describe here is related to the selection tool, not to Kiwix, neither to the storage format. Extension:Collection was able to generate ZIM files before OCG project. Help-us to bring this feature back.
IMHO best solution for this use-case would be to maintain local DB that would contain only data for pages selected by user, which would be downloaded using mediawiki API's, so that there would be no need for any ZIM packaging whatsoever.
Downloading each page separately using the API will be for sure slower than downloading the corresponding data in ZIM format... probably even if you re-download all articles (so without any incremental update process).
Having this option in Kiwix would be cool, but I can't see anything else what could be reused than the reader itself (the part of kiwix that turns wikitext into html page and display it on screen).
Kiwix has nothing to do with wikitext and never turns it in HTML.
Emmanuel
wikitech-l@lists.wikimedia.org