write a sample sTeX document in SMGloM

[kohlhase]

I want you to write a sample document in SMGloM (first one in English, then translate it into Chinese) For that you have to

• read (and discuss with me; and tell me where you need more documentation) the documentation on the main page of https://mathhub.info (and follow all the links).
• install the lmh tool for local Mathhub (but wait, we will probably have a ready-made virtual machine for that soon).
• install one of the SMGloM repositories
• read through the SMGloM documentation (SMGloM primer; http://gl.kwarc.info/smglom/manuals/raw/master/primer/primer.pdf)
• actually write an SMGloM module
• debug it with lmh pdf
• debug it with lmh omdoc
• then we can see whether we accept it into SMGloM.

[angerhang]

I think I was unable to obtain a valid MathHub account to obtain a valid copy of local SMGloM, because of some errors on the registration page assuming the link for registration is correct https://mathhub.info/smglom/smglom. First there is no place for me to enter a password and second, I can't get the "CAPTCHA" correct.

[kohlhase]

this is something @jucovschi and @m-iancu should look into

[m-iancu]

1. There is no place for a password because drupal validates the e-mail first, then asks users to set a password (that should work).
2. The CAPTCHA is MathML so it needs a MathML enabled browser to view properly (e.g. firefox) otherwise (e.g. in chrome) one sees 142 instead of 14^2. Also refreshing the link generates a new CAPTCHA but seeing it properly and roughly knowing the values of e (ℇ) and pi (π) should be enough. E.g. it should be 57 for the example below (note that both floor and ceiling could appear):

[dginev]

It is always demoralizing to fail the Turing test :)

[angerhang]

@m-iancu that is odd. The error says my password needs to be at latest 8 characters long.

1. There is no password blank for me to enter a password.
2. Based on what you said, it shouldn't give me that error either. @dginev that's very true :)

[m-iancu]

Then that is a bug, I've made an issue here for Denis: https://github.com/KWARC/MathHub/issues/108

[m-iancu]

OK, it should be fixed now. Can you try again ?

[angerhang]

Thx it is fixed. I was able to register only after several trials of CAPTCHA.

The primary reason for I had to try several times is the fact that the question is ambiguous especially when the end user sees it in a format like below (using Safari). Particularly, the get the next integer symbol of the second term is unclear. It probably would be much nicer for the user, if the human test gets a little bit more human friendly :)

[dginev]

Naturally, a "real" human being could access the XML source of the equation, read the presentation MathML code, deduce the semantics behind the symbols and then compute the final result. Though this then becomes a Turing test for KWARCies, rather than humans :)

[m-iancu]

@dginev To be honest, the out-of-the-box thinking required to post an issue describing the problem proves one to be human more than understanding a captcha. This might even be the future of anti-spam protection :).

[angerhang]

@m-iancu when I was testing the web workflow of editing my SMGloM module. I tried to access https://mathhub.info/angerhang/mySet/source/test.en.tex/.omdoc it says I don't have the permission on this server. I wonder if I have done something wrong or this happened due to some other reasons?

[angerhang]

@kohlhase I don't know if you want me to inform you about kind of error or not, but in the SMGloM primer document page 6::listing 3::the module signature code snipet::line 8. There should be no \end{modnl} as we don't need that here. It should be a typo.

[kohlhase]

yes, please report such errors, but better still, please correct them yourself :-) if you are sure that they are typos. I can give you access.

[kohlhase]

I have given both of you access to the smglom blue notes repos, you should have received e-mail.

[angerhang]

@La-Stravaganza @kohlhase It seemed to me that. Neither lmh and makefile within the repository can let me generate the pdf file successfully. When I tried using lmh pdf, the '\ref{listing}' becomes question mark instead of the actual reference number that we desire.

[kohlhase]

In looked at @La-Stravaganza's multiset, and it should have more \tref*, for instance to the concept set. Please have a look again.

[kohlhase]

looking at @angerhang's relation-wellfounded.tex, the module name restriction is an error. Also instead of \defi[relation-wellfounded]{wellfounded} you should use \defi{wellfounded}. Also you have two definitions, if they are equivalent, you should have two modules (one for each definition) and (eventually) have views between them.

But all in all things look good.

[kohlhase]

You should fix these problems before the pull requests.

[angerhang]

I fixed the problems and yet I am confused with two two modules you mentioned. (I should name these two modules differently?) As what is written in the primer, divisor module example, it also has two definitions that have the same meanings in the same module. That's why I don't understand,

[angerhang]

There is not obvious problem when running through omdoc and pdf.

[kohlhase]

If you look at the CICM 2015 paper, then it says that we associate one module with one definition (and you have two, so there should be two modules (named differently)). The two definitions in the primer you talk about the first is for "divisor" and the second one for "multiple". They are different concepts. You are right though about the second divisor definition, I have fixed this in the primer.

[angerhang]

I don't know if I have done it correctly.