Options

Introduction: Jerry Wedekind

I joined because I'm interested in John's Category Theory course. I'm a systems engineer who has worked in large printing system and transportation system development.

Probably the closest I've come to something like Category Theory (which isn't very close, as best I can tell) is to occasionally use Unified Modeling Language to help with requirements analysis or design.

The aspects of CT which most interest me at this point are its ability to tie together various domains, to represent concepts graphically, and to potentially be useful in my work.

I also hope that it can help me better learn and communicate technical info generally.

Finally, my stretch goal (we can all dream:-) would be the ability to understand 51% of what John writes on his wordpress blog.

Comments

  • 1.

    I'm sorry my blog is less than half comprehensible. But I'm glad it's at least intriguing!

    At the applied category theory workshop at NIST a couple of weeks ago, someone said that category theory should be used to develop a successor to UML. This makes a lot of sense to me, since UML seems a bit like a grab-bag of tools, but many of these tools have category theory written all over them - at least when I look at them! So, there could be a way to make UML more powerful by making it more systematic using category theory.

    Comment Source:I'm sorry my blog is less than half comprehensible. But I'm glad it's at least intriguing! At the [applied category theory workshop at NIST a couple of weeks ago](https://johncarlosbaez.wordpress.com/2018/02/17/applied-category-theory-at-nist/), someone said that category theory should be used to develop a successor to UML. This makes a lot of sense to me, since UML seems a bit like a grab-bag of tools, but many of these tools have category theory written all over them - at least when I look at them! So, there could be a way to make UML more powerful by making it more systematic using category theory.
  • 2.
    edited March 2018

    Iirc Booch described a motivation for UML being to decompose the good old flow diagram. The 14 structural and behavioural types show separate aspects such as use cases, state charts etc. (but need Object Constraint Language (OCL) text to spec. conditionals), which a tool like Rational can ensure are compatible with each other. These might be better representable by a much smaller number of different domain specific languages (DSLs)? The code and spec then form a bisimulation.

    Comment Source:Iirc Booch described a motivation for UML being to decompose the good old flow diagram. The 14 structural and behavioural types show separate aspects such as use cases, state charts etc. (but need Object Constraint Language (OCL) text to spec. conditionals), which a tool like Rational can ensure are compatible with each other. These might be better representable by a much smaller number of different domain specific languages (DSLs)? The code and spec then form a bisimulation.
  • 3.

    Zinovy Diskin, who attended the NIST workshop, has worked on the formal semantics of UML, e.g., this paper makes a connection between UML and category theory. I don't know much about this topic but I'd like to learn more.

    Comment Source:Zinovy Diskin, who attended the NIST workshop, has worked on the [formal semantics of UML](http://www.cs.toronto.edu/~zdiskin/pubs-via-interests.html#umlModeling), e.g., this [paper](http://www.cs.toronto.edu/~zdiskin/Pubs/KluwerBook-2003.pdf) makes a connection between UML and category theory. I don't know much about this topic but I'd like to learn more.
  • 4.

    I am also interested in a mapping from UML to categories. Also, SysML and two modeling languages we have used at my place of employment.

    Comment Source:I am also interested in a mapping from UML to categories. Also, SysML and two modeling languages we have used at my place of employment.
  • 5.
    edited March 2018

    John Baez how about a cats<->UML experiment some time in this course? (Or after, or on the side, or however?)

    (Hah! FInally figured out how to mention somebody effectively in this forum...)

    Comment Source:[John Baez](https://forum.azimuthproject.org/profile/17/John%20Baez) how about a cats<->UML experiment some time in this course? (Or after, or on the side, or however?) (Hah! FInally figured out how to mention somebody effectively in this forum...)
  • 6.

    Wow, this intro/chat concept really works! I'm glad to see that this is such a friendly group, and that others are interested in UML and its connection to CT. Thanks in particular to Evan for the pointer to the interesting-sounding Diskin paper.

    (John, I think you know this but I didn't mean to impugn the comprehensibility of your blog, which I think is wonderful. It's only the current-Jerry-comprehensibility which needs work - and all of the work needed is at my end:-)

    Comment Source:Wow, this intro/chat concept really works! I'm glad to see that this is such a friendly group, and that others are interested in UML and its connection to CT. Thanks in particular to Evan for the pointer to the interesting-sounding Diskin paper. (John, I think you know this but I didn't mean to impugn the comprehensibility of your blog, which I think is wonderful. It's only the current-Jerry-comprehensibility which needs work - and all of the work needed is at my end:-)
  • 7.

    Jerry, we might be working on some compatible use cases. See https://forum.azimuthproject.org/discussion/comment/16337/#Comment_16337

    Comment Source:Jerry, we might be working on some compatible use cases. See https://forum.azimuthproject.org/discussion/comment/16337/#Comment_16337
  • 8.

    Hi Bob, I did look briefly at your UML diagram and see that you're making use of associations and subclass/superclass relationships, and seem to have quite an architecture sketched out, though I didn't understand a whole lot of it for lack of context (and perhaps lack of domain knowlege):-).

    I could swear that I saw another message from you as well, but can't put my finger on it so will go back to my reading of the text for now (it turns out that I am decidedly not a speed reader when it comes to Category Theory:-)

    Comment Source:Hi Bob, I did look briefly at your UML diagram and see that you're making use of associations and subclass/superclass relationships, and seem to have quite an architecture sketched out, though I didn't understand a whole lot of it for lack of context (and perhaps lack of domain knowlege):-). I could swear that I saw another message from you as well, but can't put my finger on it so will go back to my reading of the text for now (it turns out that I am decidedly *not* a speed reader when it comes to Category Theory:-)
  • 9.

    Hi Jerry, here''s more of the context of that diagram: https://www.valueflo.ws/

    Slide deck about the model: https://speakerdeck.com/mikorizal/everything-in-valueflows-is-connected-to-everything-else

    Some of the use cases we have explored so far: https://github.com/valueflows/valueflows/tree/master/use-cases

    Re speed reading of cat theory: I am working for the slowest student award.

    Comment Source:Hi Jerry, here''s more of the context of that diagram: https://www.valueflo.ws/ Slide deck about the model: https://speakerdeck.com/mikorizal/everything-in-valueflows-is-connected-to-everything-else Some of the use cases we have explored so far: https://github.com/valueflows/valueflows/tree/master/use-cases Re speed reading of cat theory: I am working for the slowest student award.
  • 10.

    Bob wrote:

    John Baez how about a cats<->UML experiment some time in this course?

    That would be great. However, all the experts I've talked to tell me that UML is too much of a mess to neatly translate into category theory - it's a grab-bag of tools without an overarching logical organization.

    At the National Institute of Standards and Technology meeting last month, a bunch of people agreed that the successor to UML should be clearly based on category theory.

    So what I'd really like to learn UML enough to help create something else - CML? - that really takes advantage of category theory. That would be really cool - and also a lot of work.

    A first step might be to try translating UML into category theory and see what the problems are.

    Comment Source:Bob wrote: > John Baez how about a cats<->UML experiment some time in this course? That would be great. _However_, all the experts I've talked to tell me that UML is too much of a mess to neatly translate into category theory - it's a grab-bag of tools without an overarching logical organization. At the National Institute of Standards and Technology meeting last month, a bunch of people agreed that the _successor_ to UML should be clearly based on category theory. So what I'd really like to learn UML enough to help create something else - CML? - that really takes advantage of category theory. That would be really cool - and also a lot of work. A first step might be to _try_ translating UML into category theory and see what the problems are.
  • 11.
    edited April 2018

    Bob wrote:

    (Hah! FInally figured out how to mention somebody effectively in this forum...)

    Yes. Unfortunately I don't think the method you found:

    [John Baez](https://forum.azimuthproject.org/profile/17/John Baez)

    notifies me that I'm being mentioned.

    For that, I believe we'd need to use one-word usernames like JohnBaez, and do #JohnBaez. Unfortunately I only learned that after separating about 200 people's usernames. We may go back at some point... when someone with administrator privileges is really really bored and having nothing better to do.

    Another simpler method that also doesn't notify me is

    [[John Baez]]

    That's because I've taken the time to create an entry on the Wiki:

    Anyone who edits the wiki a lot should do the same. Why? Because then, when you log in and edit the wiki, people can learn who is making those edits.... which is a good thing: we're aiming for trustworthiness.

    Comment Source:Bob wrote: > (Hah! FInally figured out how to mention somebody effectively in this forum...) Yes. Unfortunately I don't think the method you found: `[John Baez](https://forum.azimuthproject.org/profile/17/John%20Baez)` notifies me that I'm being mentioned. For _that_, I believe we'd need to use one-word usernames like JohnBaez, and do #JohnBaez. Unfortunately I only learned that after separating about 200 people's usernames. We may go back at some point... when someone with administrator privileges is really really bored and having nothing better to do. Another simpler method that also doesn't notify me is `[[John Baez]]` That's because I've taken the time to create an entry on the Wiki: * [[John Baez]] Anyone who edits the wiki a lot should do the same. Why? Because then, when you log in and edit the wiki, people can learn who is making those edits.... which is a good thing: we're aiming for trustworthiness.
  • 12.

    I credited Fred Eisele with starting a lot of discussions about exercises in Seven Sketches... but then I noticed that at least one was started by Jerry here. Jerry: did you start a bunch?

    Comment Source:I credited [[Fred Eisele]] with starting a lot of discussions about exercises in _Seven Sketches_... but then I noticed that at least one was started by Jerry here. Jerry: did you start a bunch?
  • 13.

    re notifications -it's ok, I'll stop trying to mention people. They would probably be annoyed anyway...

    Comment Source:re [notifications](https://forum.azimuthproject.org/discussion/comment/16877/#Comment_16877) -it's ok, I'll stop trying to mention people. They would probably be annoyed anyway...
  • 14.
    Comment Source:NASA JPL [Semantically-Rigorous Systems Engineering Using SysML and OWL](https://trs.jpl.nasa.gov/bitstream/handle/2014/43343/12-4820_A1b.pdf?sequence=1&isAllowed=y)
Sign In or Register to comment.