Re: [Cs-studio-core] CSS BOY ideas

[Chen, X. <ch...@or...>]

Hi Tom,

>Kay came to Diamond last week to give us some training on CSS, and now I
>have some ideas about BOY to run past you. We're looking to completely
>replace EDM with BOY on beamlines, so I imagine we'll be extending the EDM
>to BOY converter, but I wanted to take this opportunity to redesign some of
>our screens instead of just converting them as they are.

That would be great if you can put someone on this effort! 

>
>* I like the way you've used the Properties pane in the OPI editor so that you
>can select elements and change their properties, and I'd like to do the same
>for certain widgets at runtime. I thought maybe make the group box
>selectable, so that when it is clicked on it loads an XML tree
>Name,PV,WidgetType,Tooltip elements that are displayed in the properties
>panel. This would mean that many of our settings screens (consisting of labels
>and text edits / displays) wouldn't have to be ported and could be
>autogenerated from tags in the database. What do you think?

Kay has discussed this with me, but we didn't get the specific use cases. Could you describe a detailed use case for me? For example, what properties should be displayed or configured? How to configure it in edit mode? How to indicate the selected widget?


>* I'd like to make a screen which allows you to rewire areaDetector plugins on
>the fly. We already have a GEF based tool in our data acquisition package that
>does something similar: http://www.dawnsci.org/tools/workflows but I
>wondered if anyone had done something like this already. If I've already
>made the properties widget I mentioned above, then clicking on one of the
>plugins could bring up its properties view to make changes to it. Has anyone
>done something similar in CSS?

No, I don't know anyone who has done workflows stuff in CSS.



>* Our GDA team use the windowbuilder pro tool for creating their GUIs:
>https://developers.google.com/java-dev-tools/wbpro/userinterface . They
>also seem to take a different approach to draw2d and SWT. Instead of
>creating lots of draw2d and SWT widgets, and placing them on one big draw2d
>canvas, they create mostly SWT composites, with a few draw2d widgets on
>their own canvases, and use windowbuilder pro to lay them out using SWT
>layouts. Is this an approach that you've considered?

I considered something similar: a composite grouping container or linking container, in which case, the container is a SWT composite. The only benefit I can think of is for WebOPI: it can limit the canvas redraw inside the composite which may improve the performance. But this requires too much work and could also make user confused with too many similar functions widgets as Kay mentioned. Do you see any other benefits of it?

Thanks,
Xihui