Re: Causeway viewing your E-on-CL traces

Kevin Reid <[email protected]>
Newsgroups gmane.comp.lang.e.general
Message-ID <[email protected]>
On Aug 5, 2008, at 18:47, Tyler Close wrote:
> On Tue, Aug 5, 2008 at 1:42 PM, Tyler Close <[email protected]>  
> wrote:
>> On Tue, Aug 5, 2008 at 11:38 AM, Kevin Reid <[email protected]> wrote:
>>> If they are to be generated relative, what should they be relative  
>>> to?
>>
>> Assume all source code is found under some root directory. The source
>> path is then the relative path from this root to the source file. ...
>>
>> "source" : "example/src/org/waterken/bang/Drum.java"
>>
>> which assumes a root folder under which each project is stored.
>>
>> Obviously this means all the source code in a given interaction must
>> coordinate their chosen project names.
> ...
> The contention on the project name is unfortunate, given that part of
> the path name already encodes a globally unique identifier. It might
> make more sense to make the path be:
>
> "source" : "org/waterken/bang/Drum.java" ...
> I think this would do a better job of decoupling source packages.  
> Thoughts?

I like this approach, with the caveat that if I get my way, E will  
move away from depending on the DNS system for module isolation, and  
I'm not yet sure whether the resulting system will have globally  
unique identifiers *in the pathnames*.

Also, there will always be source files which are not organized unique- 
name-wise: top-level programs ("scripts"), caplets, etc. I don't see  
how to generate reasonable relative paths for these.

On Aug 5, 2008, at 16:42, Tyler Close wrote:

>> Perhaps the log should have a field for a nickname (in the Zooko's-
>> Triangle sense) for the vat, since one exists in E and it's more
>> likely to be what the user wants?
>
> The event loop URIs in the Waterken server take the form:
>
> "loop" : "http://localhost:8080/-/t7/",
>
> where everything after the "-/" segment may be a programmer chosen
> name. I figured the Causeway viewer could use the last segment in the
> URI as the short name and the full URI as the globally unique name.


I think this is a decent and simple solution. It Violates The  
Opaqueness Of URIs, but has lots of precedent. I'll do it.

-- 
Kevin Reid                            <http://homepage.mac.com/kpreid/>
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.