Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
Scott Morrison has a nice citation plugin for MO. The search engine already seems quite usable, but I am wondering whether there are more/better ways of integrating this with the nlab. I though, I’d mention it.
That’s interesting. I couldn’t readily find what DB(s) they are searching to find references.
I don’t particularly care for their meta-data/micro-format representation but there is no common standard, the whole field is fragmented, and people who are really into standards create such huge things as Medical Scholarly Article.
The nLab, being more of an encyclopedia than a blog where you just need a quick way to cite something, should have its own DB of papers/books cited with citations on pages handled by inclusions. This would allow nlab citations to have a consistent style (such as linked full author names, arXiv track backs, and what else).
Under this scenario, to create a cite in nLab you type in enough text to identify a reference and then the nLab searches its DB to see if it already exists. If it does then it returns something like [include /refs/mac037] which the user can modify to be something like [include #fullref /refs/mac037] for something at the bottom of the page, or [include #citeref /refs/mac037]for a tiny inline reference that links to a #fullref or even pops up the full text on mouse hover.
If the reference is not found in the nLab DB then it uses the MO search or similar and uses the returned data to guide the user in adding a new entry. I would hope the data includes both BibTex and RIS (if available) because sometimes one is better than the other.
The nLab, being more of an encyclopedia than a blog where you just need a quick way to cite something, should have its own DB of papers/books cited with citations on pages handled by inclusions. This would allow nlab citations to have a consistent style
nLab is not an army unit and total uniforminess is not a must, nor even desirable. The same reference is in some article cited by link to a reference page, or is just given a short format as the reference is nt that important or the reference list is long. If the reference is very pertinent to some article then t may be useful to give MR number, doi, link to arXiv, link to free pdf, link to english translation (as e.g. Tohoku), link to Bull. AMS review article (for a book), link to Errata etc. The reference can have an anchor to refer to it from within the entry, Sometimes, if the reference is important and not accessible online, one can include a part of the abstract. Not always we need all of that. Also if one author s an author of 4-5 related references we can write them with quoting the authors name only once and chain the titles one after another without opening a new item line. This makes it more compact. One has to weigh between putting much data and being compact and easy to scroll and notice the author or subtheme groups.
I also should point out that some groups of entries are dominantly written by some groups of authors, and each group may have their own preference in formatting. Also notice that some non/English authors do not have the same spelling in each of the published papers.
This is not, of course, against having a database, but against automatic overwriting of hand-input references. Hand-input stuff is the most precious.
We have discussed having a citation database before; I'll try to find the thread (or threads). It would be nice to have a convenient and flexible way to cite things.
I feel like there was something yet later too, but I can't find it.
Elsewhere with Adeel we are experimenting with WorkingWiki. That supports bibtex and allows to have a .bib database from which to include citations.
Zoran, to my mind the standardization of typesetting in mathematics (of definitions, theorems, etc, but also of citations) serves a major practical purpose: it eases the trouble that the reader has with extracting information. The more standardized the text outline is, the quicker can I scan it for relevant information.
Think of it this way: there is such a gigantic need and room for creativity and mindpower in mathematics, that we don’t want any of that to be wasted on what should be plain mechanical tasks, such as extracting the page number of a citation.
we don’t want any of that to be wasted on what should be plain mechanical tasks
The new tools are welcome! But you seem also here defending and suggesting overwriting hand-written data what is not welcome. If the data is not yet written generate it from Alpha-Centauri’s machines or whatever, but why to spend energy overwriting (in that case many specific information would be overwritten); you could just add a link to standard format externally if you like it. The main reader is the AUTHOR of the edit, you were advertising us to make the things for our own use (“selfishness”, which I was less eager about), so this would annihilate anything what is tailored to specific needs or tastes of authors of specific groups of entries.
Besides, finding specific data sometimes needed like Bull AMS review articles, Russian translations, alternative depositories, free excerpts and so on is not such a plain mechanical task. It is not trivial what to omit (which parts of the complete bibliographic data) in some entry (weather the reference is central or not and who is the group of supposed users, what is different for simple entries or from more specialized ones) and if you propose to omit nothing than the lists of literature will be huge graphically, what is not very practical and it does slow us down.
I am just saying that standardized formatting in math is – contrary to practices in the army that you compared it to – a means to increase creativity instead of decrease it. Namely the more mechanical it is for the eye to find information in the text, the less is one distracted from what really matters.
Namely the more mechanical it is for the eye to find information in the text, the less is one distracted from what really matters.
This is abstract idealization which does not answer any of the concerns above (is the list above so incomprehensible ?). Do you include everything (all available information on some reference is displayed in every single page which includes it into its bibliography) ? If yes the page output will be typically so huge that it will be hard to scan for any eye. Do you want to take bare minimum ? Then specific info some entries or users need will not be put in the moments when it is good to include. So how will you make good choices if you propose overwriting hand written info in a uniform way ? You may have something in mind what I am not aware of. Please let us know if so.
Also, will the information be in the source of the page, or only outside ? (we can have two types of data in fact simultaneously) For me it is important to have an easy possibility that the source of the page has the references inside, as I often save just a page (and its source!) I need for daily usage offline. Maybe you did not explain what you actually plan to do (without abstract phrases on the eye – mechanics how would you like things to work and info to flow).
Also one should have in mind that most of us do not know and never used BibTeX-tools while all of us use plain LaTeX with bibliography list, so the rest of us should be able to use it without ever recoursing more than trivial to BibTeX-specific knowledge and commands (most particularly when cutting and pasting back and forth between our personal LaTeX files and Lab sources).
For future reference, zbmath has a new tool for finding citations.
1 to 10 of 10