There are many sites that contain hebrew texts for free, such as http://he.wikisource.org, http://mechon-mamre.org and http://hebrewbooks.org. Yet it is often inconvenient to go to the sites to access the texts. Would there be some way that one could just refer to the texts, and they would automatically become links? What about having a built-in search of basic Torah texts?

  • 1
    MediaWiki has such a thing built in; the version used on the Wikimedia sites has a whole slew of sites linked to by abbreviations so that, for example, [[arxiv:math/0407161]] creates a link to http://arxiv.org/abs/math/0407161. However, this is really a feature request for meta.SO; I suggest we move this discussion there.... Oh, wait: it's already there. (This is all re the first suggestion, that linking be easier.[continued]
    – msh210 Mod
    Commented May 24, 2011 at 3:26
  • [continued] A search, the second suggestion, sounds to me like something that's Not Gonna Happen, but that's a guess.)
    – msh210 Mod
    Commented May 24, 2011 at 3:29
  • 1
    I like the idea of a special auto-tag feature a lot. If it only provided for Biblical verses, it would already be very useful. If we could develop and open-source a markdown extension to do it, it could be incorporated here (and on at least one other coming SE, lehavdil) and a great gift to the world.
    – Isaac Moses Mod
    Commented May 24, 2011 at 14:32
  • ... or even a one-stop web 2.0 tool that generates link code based on selections from dropdown menu. For example, the first menu would let you choose the corpus (e.g. Bible or Talmud), the second would let you choose the book (e.g. Orach Chaim or Yoreh Deya), the third would let you choose the chapter, and the fourth would let you choose the verse/paragraph, if applicable. I'm tempted to learn AJAX techniques just to build such a tool, since I think it would come in handy to many people.
    – Isaac Moses Mod
    Commented May 24, 2011 at 15:05

1 Answer 1


Update: I edited my post to be a little more clear...

From a technical standpoint:

Unless we introduce something in markdown to specifically "tag" stuff as a reference, it'll be a real pain. Not saying impossible, just saying very difficult to implement properly.

Suppose we decide to go with [ref:SeferName (numerical:hierarchy:of:identifiers)], which is a great idea (thanks @msh210). We would need a script that magically:

  • Identifies SeferName in a standardized way (think: MishnahBerurah, MishnaBrura, MB.....)
  • Associates URL with SeferName
  • Figures out correct hierarchy of identifiers with said SeferName
  • Figures out correct URL (or in case of PDF's page numbers) of said Hierarchy
  • Oh, and maybe page numbers and paragraphs too?

Obviously, this would require having a reference database, as mentioned in the comments.

I have no problem creating such a database. But it would require a bit of a learning curve: both in terms of filling the database with entries (anyone else volunteer?), and in terms of usage. It would also have to be super-carefully though out and highly standardized. Something to be discussed thoroughly...

And convincing SE to implement it afterwards... Well, that's something we would have to discuss with them. If anything we can create a userscript (for editors) that'll do it, but....

And the benefit is?

The real question is what do we gain? Obviously having this makes life easier for editors, but does it have any other benefits?

Essentially, the best way to reference is by providing all the info about the location in the sefer in plaintext and letting the future user look it up either in print or online. This does provide a shortcut, but at what expense?

From an implementation standpoint:

Which site should we use? You pointed out three. Wikitext probably has the most potential for staying around for years to come ($$$), and already sort of has a referencing type system (where names of Seforim and numbers are all included in the URL in a uniform way).

On the other hand, HebrewBooks has the largest selection (no doubt), but uses cryptic "Id"-like numbers for all it's seforim. Which goes back to what we said above about the database.

Bottom Line

I think the most practical solution for now would be to simply refer to the text, if you're feeling particularly lazy or are in a rush. Ideally someone else will come around and edit in the link.

As for a more long-term solution. Working on a database that will take care of all problems discussed, in a reasonable, permanent, standardized, backward-compatible, and future-proof fashion.

  • 2
    I agree with the bottom line. However, a way to tag things shouldn't be impossible. We already have the [tag:foo] syntax, so why can't it be extended to [ref:MishnaBrura:698.1]?
    – msh210 Mod
    Commented May 24, 2011 at 3:55
  • 2
    Use of a userscript would depend on each visitor to the site's having the script installed, so we couldn't do that. Unless I'm missing something.
    – msh210 Mod
    Commented May 24, 2011 at 3:56
  • @msh210 B) We can have a userscript for editors that'll make the process easier / automatic. A) Like I said, it would need to be in "tag" format (not exactly automatic), and would need to be very forgiving in terms of spelling, punctuation, and transliteration. Plus it would need a database of source materials with corresponding source URLs. And is that really easier than finding said text in wikisource and copy-pasting link?
    – yydl
    Commented May 24, 2011 at 4:22
  • A) It wouldn't need to be forgiving: it could require a certain spelling, and people would look up that spelling on a master list until they knew it. It's doable: it's done (in the WM sites; see my comment on the question). If you are starting with a paper text (or memory) then, yes, that will be much easier IMO than then digging up the digital copy and linking. B) Ah, that sounds good. Some of the SO users on this site might be able to help with that (using jQuery, of course).
    – msh210 Mod
    Commented May 24, 2011 at 4:35
  • A very rough draft (since I am not a coder): var v=document.getElementById('wmd-input').value;v.onchange=if(v.match(/\[ref\:MB[0-9]+\.[0-9]*\]/))v=v.replace(/\[ref\:MB([0-9]+)\.([0-9]*)\]/g,"something");
    – msh210 Mod
    Commented May 24, 2011 at 4:51
  • @msh210 while that's definitely not valid JS (there is no onchange for value, only for element itself), you hit my point. What is this "something" you are referencing? Without a database that maps References -> URLs, there is no way to go about this.
    – yydl
    Commented May 24, 2011 at 5:09
  • Well, if the JS is along the lines of what I wrote (but valid ;-)), then it would presumably have built into it (or call) such a database. That's simple (if time-consuming) to do manually.
    – msh210 Mod
    Commented May 24, 2011 at 5:37
  • function MishnaBruraVolumePage(siman,sif){switch (siman*1000+sif){case 1001:case 1002:case 1003:return 1001;//etc etc etc}};function MishnaBruraURL(volumepage){switch ((volumepage-(volumepage%1000))/1000){//find the right book on hebrewbooks};//etc}
    – msh210 Mod
    Commented May 24, 2011 at 5:48
  • There is precedence, once the site leaves beta, to replace the "Code Sample" formatting function. Since we have (almost) no use for that here, it would definitely make sense to petition SEI to build up some form of auto-reference engine/format to replace the Code Sample. The only difficulty would be in deciding on the "authoritative" source.
    – AviD
    Commented May 24, 2011 at 6:49
  • 1
    @AviD Could you please point to some examples of this function replacement?
    – Isaac Moses Mod
    Commented May 24, 2011 at 15:16
  • @yydl What about, as I suggest in the Q comments, a menu-based, independent web tool as front-end for the database instead of a tagging system? That would eliminate your first issue (spelling), would not require SE cooperation, would be useful for people all over the web, and could potentially offer options of different sites to prefer getting links from.
    – Isaac Moses Mod
    Commented May 24, 2011 at 19:18
  • 1
    @IssacMoses Exactly what I had in mind by database. A central location to refer to, which then can direct to other sites at will. Have some ideas on how it should work, maybe will get a prototype out soon, but really wanna see how this discussion forms, for better ideas on how it should work.
    – yydl
    Commented May 24, 2011 at 19:22
  • 1
    @yydl If you are serious about starting this project and could stand collaboration, I'd love to help. I don't have a great deal of web programming experience, but I've done plenty of other programming and HTML editing, and I have some ideas about how to make this tool awesome.
    – Isaac Moses Mod
    Commented May 24, 2011 at 19:33
  • 1
    @IssacMoses Yes, serious. Collaboration is a must, not an option. Very good with HTML, CSS, JavaScript, AJAX, Python, etc. so that shouldn't be a barrier. The real barrier is guidelines: implementation needs to be discussed and dissected thoroughly before anything concrete is put into place. Not sure what the best way to do that would be, but we've already got a lot of it on this post...
    – yydl
    Commented May 24, 2011 at 20:29
  • 1
    @Isaac, there is LaTeX support on Tex - LaTeX, MathJaX on Math.SE and some others, and "spoiler" tags on Gaming and Science-Fiction. And that's just the ones I know about...
    – AviD
    Commented May 25, 2011 at 10:19

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .