[Grok-dev] Re: Admin UI name change suggestion

Philipp von Weitershausen philipp at weitershausen.de
Fri Oct 5 06:33:06 EDT 2007


Martijn Faassen wrote:
> Philipp von Weitershausen wrote:
>> Martijn Faassen wrote:
> [snip]
>>> If you turn the Zope 3 root into a grok application, there are some 
>>> questions, like:
>>>
>>> How do you remove the application again and reinstall it? This is 
>>> what I do quite a frequently during development.
>>
>> As I've suggested in another reply, this could easily be handled by a 
>> view.
>>
>>> Where is the user interface to go to the introspector?
>>
>> Append /introspector or something? I don't think it's too hard to come 
>> up with an easy-to-remember rule how to reach it.
> 
> So you're proposing we *throw out* the UI entirely in favor of various 
> URL appendages we need to remember?

No. I'm proposing that you get to the UI not by saying 
http://localhost:8080/ but by saying http://localhost:8080/introspector.

I don't understand where you got the idea I want to throw things out. 
All I talked about was a default choice that Grok makes for you.

>> Less work to get started. Less concepts to grasp (object publishing, 
>> sites as in ISIte, etc.).
> 
> It depends on what you mean by "getting started", right? It's less work 
> to get a single application installed.

Please convince me that this *isn't* what 99% of the people starting 
with Grok want and I'll be happy to reconsider my view :).

> It's more work to figure out everything else. Lack of a UI will make 
> life harder to get started.

Assuming people want this UI.

> You don't know how to get to the introspector (you need to read this somewhere). You don't know how to 
> recreate the application (you need to read this somewhere). You can't 
> detect that your application is broken (again, you need to read this 
> somewhere).

 From my experience from working with newbies, they don't detect the 
"broken" state anyway. They wonder why their new utility isn't showing 
up, or why they get a NotFound. I *always* had to explicitly tell them 
that they had to reinstall applications. When I did so, it made sense to 
them, but they too had to inform themselves about that fact.

I don't think anybody ever intuitively turned to the admin UI for this. 
They were simply refreshing their browser pages, and found that they 
stopped working.

> I don't understand how there are less concepts to grasp. People working 
> with Grok will need to understand object publishing;

Yes. Eventually. I distinctly remember in the discussions around the 
first prototypes of Grok that object publishing was not the first thing 
people should have to learn

> do you think this will somehow avoid this?

No, but postpone.

> I also have the strong suspicion that 
> eventually people will encounter the ISite concept as well.

Sure. Grok certainly makes that easier.

>>> So far I've heard "people are confused", but it's not clear to me 
>>> that this change would make people less confused or if so, it would 
>>> actually improve their lives?
>>
>> Well, everybody who said they were confused said that they would've 
>> not expected this step. That this application object "would just be 
>> there". That seems pretty clear to me about whether it would improve 
>> their lives.
> 
> One wouldn't think they would be confused for long as soon as they go to 
> localhost:8080, right?

I don't understand this question.

> Anyway, again, having a way to get an application installed by default 
> would be useful in various circumstances.

WOuld you agree it made sense that one of those "various circumstances" 
was the default setup that grokproject produced?

> It just doesn't seem to be the driving use case to:
> 
> * throw out the admin UI (we should make it optional, of course)

not what I'm proposing

> * make installing multiple applications an "expert" use case only.

perhaps "expert" was the wrong word, but I certainly think it's more of 
a second-week-with-grok use case than a first-day-with-grok one ;).

> * generally do a lot of work to figure out all kinds of use cases that 
> we already cover today.

I think they could still be easily covered.


-- 
http://worldcookery.com -- Professional Zope documentation and training


More information about the Grok-dev mailing list