[Attempto] Workflow management and no-code apps creation

[Kaarel Kaljurand]

Hi,

answering some questions which were maybe not answered yet:

I think it makes sense to install the commandline version of APE
(https://github.com/Attempto/APE). It has at least the following
benefits:

- you can specify your (base) user lexicon at compile time, which is a
bit faster (depending on the size of the lexicon)
- you can run an HTTP server locally and use it from your favourite
programing language however your like (there is special support for
Prolog and Java, of course), rather than playing with the web
interface
- using the Protege plugin (I assume you're using
https://github.com/Kaljurand/aceview/releases/tag/v2.0.0-alpha,
because older versions do not work with the latest Protege) means that
you are also getting OWL-related warnings/errors (because only a
subset of ACE is supported by the plugin), which might be confusing

Regarding modelling, you could also try OWL-based modelling (and the
corresponding ACE subset). That means that everything must be a class,
2-ary property, or an individual, e.g.:

- Every lesson contains at least 1 task. (i.e. "lesson" and "task" are
classes, i.e. sets of individuals, and "contains" is a property)
- Lesson1 is a lesson. ("Lesson1" is an individual, here a member of
the class "lesson")

It also means that you need to be explicit about Open World Assumption
and Unique Names Assumption. For example the statement "A task should
have status – ‘not started’, ‘started’ or ‘completed’." must be made
more explicit which ends up being quite verbose:

Every task contains exactly 1 task-status.
No task is a task-status.
Every task-status is NotStarted or is Started or is Completed.
If something is NotStarted or is Started or is Completed then it is a
task-status.
NotStarted is not Started.
Started is not Completed.
Completed is not NotStarted.

Of course some of the complexity can be hidden from the end user (e.g.
AceWiki asserts the uniqueness of individuals behind the scenes).

In the end what is important is what kind of reasoning (or querying)
services do you want to offer to the end user.

Best,
Kaarel

On Mon, Dec 28, 2020 at 9:37 PM Pierre-Alexandre Voye
<ontologiae at gmail.com> wrote:
>
> Norbert,
>
> I answer inside the email
>
> Le sam. 26 déc. 2020 à 08:58, Norbert E. Fuchs <fuchs at ifi.uzh.ch> a écrit :
>>
>>
>>
>> > On 26 Dec 2020, at 24:09 , Pierre-Alexandre Voye <ontologiae at gmail.com> wrote:
>> >
>> > ... "Lessons" is an application. ...
>>
>>
>> Pierre-Alexandre Voye
>>
>> Two comments.
>>
>> First I do not understand why do you define the word lessons as the string "Lessons", i.e. as a sequence of Unicode characters? Why not define it simply as a noun? In this case your sentence should probably read
>
> I did it because in the examples Igor gave us, he quoted Lessons, so I stupidely reproduced...
>
>>
>>
>> Every lesson is an application.
>>
>> Note that other than Igor I introduce lessons as a universally quantified noun.
>>
>>
>> Second, of course you can use the terminal command
>>
>> $ ./ape.exe -ulexfile lexicon/clex_lexicon.pl -cdrs  -cparaphrase1
>>
>> but the output you are looking for is much easier to be seen when you use the web-interface of APE at http://attempto.ifi.uzh.ch/site/resources/.
>
>
> As a nerd, I always prefer a terminal ;-) I use it much as possible, it's quicker : in 1s I have the good directory and the command, it would take 10s to find the web interface via google
>>
>>
>> Regards.
>>
>>    --- nef
>
> _______________________________________________
> attempto mailing list
> attempto at lists.ifi.uzh.ch
> https://lists.ifi.uzh.ch/listinfo/attempto