[Return]
Archived Thread

File 130054708784.png - (1.61MB , 971x1359 , fc33dcf30a959a6a7d454e683a3a0914.png ) [iqdb]
11775 No. 11775
This has nothing to do with eating babies

It is my personal opinion that the current methods of distributing and storing Touhou doujinshi are woefully lacking. On imageboards (here, /t-h/, the Chinese Komica board) releases are uploaded to mediafire or another fileshare and the links are copied to the . Aggregator sites such as e-hentai are restricted to members and require credit to download (or so I've heard; I honestly don't use e-hentai) and are riddled with ads. A handful of blogs exist (e.g. http://touhou-hentai.blogspot.com/) that largely work the same as the imageboard, but have all of the constrictions of the regular blog format. And there is also Mysterio006's doujin spreadsheet (more-or-less mirrored on the /hdoujins page), and the ever-useful Doujinshi DB that provides information but not downloads.

The main problem is the lack of cohesion between actual releases (e.g. the download links on mediafire/e-hentai) and the doujinshi themselves. Mediafire links provide no information about the doujinshi beyond the name, circle, and release date, and even that is only if the filename is tagged correctly. The proposal is for the creation of a new site, one specifically made for storing and indexing Touhou hentai doujinshi. The site would effectively be a merge of the information contained within the Doujinshi DB with the full-archive downloads available from imageboards.

What I would consider to be ideal would be a site that lists doujinshi, rather than directly showing the releases. Searching for a doujin by name, either Japanese or English, should both bring up the same doujin entry, giving all known information about the doujin (release date, circle, artist, all possible names, etc). The page would also list all available releases on the one page; this would mean that there is never any confusion about whether or not a doujin is translated because if it is translated it is available next to the RAW. Listing all releases could also include none, so a well-anticipated but unscanned doujin could have an entry with no releases, and anyone looking for that doujin would find the information page that also states that nothing has been released yet. People don't search for releases; they search for titles and circles and are given releases, or nothing if the doujin hasn't been released.

Examples: searching for “Cirno Showed Up at My House!”or “チルノがおうちにやってきた!” would both bring up the page for Nounai Kanojo's work. The page would contain all relevant information about the doujin itself, e.g. name, author, artist, number of pages. The translated and RAW download details would also be included, giving all relevant information about each release, e.g. type (RAW/translated, decensored/censored, quality of scans, and combinations of these), filesize, translator if applicable, and of course the download link or hash. Searching for “A Night So Hot It Freezes”, “A Hot Yet Cold Night”, “凍える程熱い夜”, or “Kogoeru Hodo Atsui Yoru” would bring up 04U's work, including the RAW, the CGRascal translation, and the retranslation by 元祖のプラスファグ. Searching for “諏訪子様と神奈子様の仰る通りに!” will bring up the info page but with no releases, as it remains unscanned.

Searching by circle, artist, convention, characters, fetishes, translator, etc would also be possible, much like on the Doujinshi DB. Constructing a real site to index doujinshi would allow functionality to be built directly to suit doujinshi, rather than hijacking existing functionality and trying to coerce it into working in a different way. Other useful additions, such as a tag system, content descriptions, comments, and contributions to doujinshi information from all users and not just the uploader could all be created as well.

Actually managing uploads is a bit of a problem though. Storing all archive files on the server and sending them to clients will naturally eat up a lot of storage and bandwidth (my mediafire reports 38.3 GB of content and 9.8 TB of transfers, however that also includes doujin games), plus putting files directly on the server would run the risk of DMCA takedown notices (Of which mediafire has mentioned two). While being able to simply upload files to the server and let everyone download freely would be ideal, the costs of doing this would likely be prohibitive. I believe that the best compromise is to set up a private P2P network that would contain all of the releases and have the website contain release notices and hashes. I've never used it, but I believe that something such as Gnutella would work well. There would be one master node containing all of the releases to guarantee availability, and the users would upload as well to balance the load. Share hashes could also be included for RAW releases, but as this isn't possible for translated releases they should only be used as a supplement to the main distribution method. Simply continuing to store releasing on mediafire is also possible, but has the same potential issues for DMCA takedown notices and would also limit the amount of automation that could be done upon uploading.

I called this a proposal because that is precisely what it is. This is just an idea I've had; I use PHP at work but I'm certainly not good enough to design an entire site, the database, page layout, and also make it secure against any hacks beyond SQL injection. So if anyone has anything to contribute, be it your time, expertise, suggestions, criticisms, anything, please do so. This is just an idea, and for a fairly complicated site, so it won't be created without plenty of support.

No. 11776
File 130054952228.jpg - (57.12KB , 844x241 , 10 seconds in paint.jpg ) [iqdb]
11776
tl;dr system is fine, change sucks, and good luck with the funding of a site that would be running under extreme traffic.
No. 11778
Filehosts are no problem as long as a file is in multiple places. Internet == Natural Backup. Indexing is a problem, but I prefer ad-hoc methods like the spreadsheet.
No. 11785
Why so pixel, Okuu?
No. 11786
>>11785
The original was too big, and I guess the resize fucked up.
No. 11798
As far as a database is concerned, you can probably just copy the format from something like vndb or mangaupdates or something. If you insist on providing content, p2p would be the cheapest bandwidth-wise but you could also always have an infrastructure solution like bots on irc. The downside to the latter is cost and honestly it tends to skyrocket for these sorts of things.
No. 11802
>>11798
P2P was suggested precisely because it should be the cheapest bandwidth wise. The users would share the load, and if the traffic becomes too much to handle older releases could be removed from the server-node, and they'd still be available for download from peers. I don't think IRC bots would be much better than direct downloads, since unless other people donate bots all of the traffic would still be on the server.