Yeah, I like the idea of an "advanced" label for non-common options. I've
often been overwhelmed myself at the list of options for some components
:-) I imagine it also clutters up UIs for the sake of showing options only
a small number of people may ever use.

Cheers,
Jon

On Wed, Sep 9, 2015 at 2:23 PM, Daniel Kulp <dk...@apache.org> wrote:

> I’m working with the UI folks in Talend to enhance the tooling that we
> provide.   Right now, we have various components with specific parameters
> exposed.  This “kind of works”, but as new parameters are added in Camel we
> have to keep things up to date and such.    We’re trying to switch to using
> the json stuff that is generated at build time (great stuff, BTW) so that
> all the parameters that are available is presented to the users.
>  Currently, there are labels for “consumer” and “producer” and anything not
> labelled one of those is common between the two.   That’s a great start.
>  However, while testing with some users, we’ve run into a usability
> problem:  some of the components have soooooo many parameters, many of
> which would rarely ever be used, that trying to find the useful stuff that
> they care about is harder than it should be.   For example, take a look at
> the file component:
>
> http://camel.apache.org/file2.html
>
> There are 7 “common” options, but for the consumer, there are then an
> additional 51 options.  It’s really too many to present in a single list
> and not have the user go “what?!?”.
>
>
> Would there be any objections to adding an “advanced” label to options
> that are generally not normally used by say 90% of the use cases?   I know
> that determining whether something is “advanced” or not can start heading
> down the path of bike shedding and the whole “what’s advanced for one user
> isn’t for another” and all that, but I think we need to have something to
> help out with this.   Another thought was an additional  “advanced=true” or
> “userlevel=1” or similar attribute to denote this, but the label
> functionality is already there and it seems to be perfect for this.
>
> I did quick look to see how many options we have.   I parsed/processed all
> the component json files (didn’t look at the models or data format) and
> there are over 10,000 options.   It’s a huge task to go through each of
> them, but if we can agree on a path forward, we can start tackling some of
> the more used components and then see how that works.   File, http, jetty,
> jms/sjms, etc…    Longer term, this info could be used to provide more
> useful component pages for the docs, putting the more useful stuff at the
> top.
>
>
> Thoughts?  Objections?  Other ideas?
>
> --
> Daniel Kulp
> dk...@apache.org - http://dankulp.com/blog
> Talend Community Coder - http://coders.talend.com
>
>


-- 
Cheers,
Jon
---------------
Red Hat, Inc.
Email: jans...@redhat.com
Web: http://redhat.com
Twitter: jon_anstey
Blog: http://janstey.blogspot.com
Author of Camel in Action: http://manning.com/ibsen

Reply via email to