[Logiweb] Curious Logiwiki behavior

[Robert Lamar]

Klaus,

Klaus Ebbe Grue wrote:
 > Robert Lamar wrote:
>> I am a PhD student in Edinburgh, Scotland who was first introduced to 
>> Logiweb at MKM '07 [1].  Just this month, I have finally found time to 
>> start investigating it.
> 
> Welcome to the list

Thank you.  :c)  I appreciate the welcome and the prompt responses to my 
question.

>> [ ... description of my problem... ]
>> ---
>> File Form input around line 153 char 16:
>> No interpretations
>> Goodbye
> 
> The "No interpretations, Goodbye" error message is one of the most 
> common error messages from the system, and it can be quite frustrating. 

In fact, this was one of the main things I wanted to confirm: that it 
was "to be expected".  Even if it is unhelpful, it's good to know that 
all users cope with this at some point.

> [ ...analysis of the specific error... ]
> 
> The 3.2l construct is one you are supposed to introduce. To do so, add a 
> line saying
>   "" 3.2l
> in the first PREASSOCIATIVE section in the upper window on the 
> submission form. 

I recall this from the tutorial, and my future (successful) attempts 
included the line in question.  This may have been the primary reason 
for confusion:

   1. On one attempt, I would overlook that line, and it would fail.
   2. Because I was working on the crossbrowser submission form, I could
      not go back to my code to find errors.
   3. On the following attempt, I was more attentive, and did not make
      such mistakes.

I have learned my lesson, and now make my experiments in a text editor, 
copying them from there into the submission page.

> [ ...more problem analysis... ]
> 
> The 3.2l line comprises an empty string followed by the 3.2l construct. 
> The empty string indicates that the 3.2l construct is not imported from 
> any referenced page but, rather, is a construct you define. Tutorial T05 
> gives a systematic treatment of how to introduce constructs.

Thank you for the thorough explanation.  I will reach Tutorial T05 soon, 
and look forward to gaining further insights.

> I think I will just give answers to two obvious questions right ahead:
> 
> Q: Why doesn't the compiler give a better error message?
> 
> A: It can't. Compilers for languages with a fixed syntax can typically 
> give very precise error messages like "semicolon expected" or "3.2l 
> unknown". But the present language has no fixed syntax and that makes it 
> *extremely* difficult to give a better error message then the above when 
> something is unparsable. Fortunately, the system can give better error 
> messages in other situations. See Tutorial T05 for more on the syntax.

This is reasonable.  I can see that improving the error messages is a 
hard problem, and should be able to live with the situation.

> Q: Then, why doesn't the website or the tutorial give an explanation 
> like the one above.
> 
> A: That is because of lack of time. Sorry.

But of course.  Any documentation is appreciated, but it certainly takes 
time and energy to produce.

Thanks again for the prompt and thorough answer.  I may be dropping in 
again with other questions, but my previous query has been satisfied.

Cheers,
Robert