Re: My Workbench wish-list
Hi!
max goldberg Wrote:
-------------------------------------------------------
> Hello, I've been using (or at least trying)
> Workbench 5.0.12 (both regular and SE) for the
> last month or so, and I have a large amount of
> feedback. Rather than clogging the bug system with
> possible bugs and feature requests, I figured I'd
> post here and give an idea of what I'd like to see
> change in Workbench in the future.
There's no such thing as "clogging the bug system" :) So
feel free to post feature requests and bug reports as you
find them.
But discussing feature requests and ideas in the forum is
good too.
[...]
> or impossible to use. So in summary, PLEASE focus
> on the general usability and intuitiveness of
> Workbench, I think it's the area that is lacking
> attention the most. If you try out any number of
> other DB design programs, you can hit the ground
> running and design a fairly simple database
> without any confusion. This should be a goal.
You're right, there are still many rough corners in the tool,
but we wanted to get the basic functionality complete
before moving to polishing the overall usability. Making
everything comfortable and smooth takes its time, but
we're working on that.
>
> Major features
> -----------------------
>
> * Schema versioning and automatic version
> incrementation.
> -- The ability to say "This is version X of the
> schema" and then later go back to
> -- previous versions of the schema.
> --
> -- An option to automatically increment the
> version number upon saving would be nice.
This is a planned feature for some future version.
We're thinking on a cvs like version control of the model
(save state of the whole model as one version) and perhaps
also saving the entire undo history to allow you to undo
across sessions. Storing in a database and tracking changes
per user is also planned.
>
> * Field recycling of some sort.
[...]
> -- MicroOLAP's database designer have "domains"
> where you can define a field
> -- and all of its properties and then when you are
> designing a table you can
> -- just select a dropdown and have that field
> inserted into your schema.
This is a feature that is also planned for the future.
>
> * Smart auto-layout
> -- When working with large schemas with 60+
> tables, (especially on a new import),
> -- auto-layout ends up with a lot of overlapping
> tables and references, requiring
> -- a great deal of effort from the user to move
> things around just to make the
> -- schema readable. It would be incredibly nice to
> have a "deep" auto-layout that
> -- will automatically enlarge the diagram size and
> then arrange all of the tables
> -- as best it can so that there is no overlap.
[...]
Auto-layouting is a tricky issue. The current implementation
has problems and we'll fix what we can, but since there is so much
to do in different areas of the project and different algorithms
can work better for different types of models, we'll also leave it
open to the community to write additional auto-layout plugins.
> Minor features
> ------------------------
>
> * Auto-save
This is in the TODO list for soon.
>
> * Model "Object defaults"
> + Documentation, or at least a list of the
> various %variables% on the options page would
> be helpful.
>
We'll add a tooltip or some label with the list of those.
> + Singular versions of table and column names,
> e.g. table "servers" would have a singular of
> "server"
> e.g. servers.server_id, services.service_id,
> server_services.server_id,
> server_services.service_id is a popular naming
> style.
> - you could go as far as "copies" -> "copy",
> but that may be more semantic work than it's
> worth.
Another problem is that different languages have
different rules for "pluralizing" words, but perhaps
a list of customizable regular expressions would work
well enough (eg: (.*)ies -> \1y, (.*)s$ -> \1)
I think this is not that urgent, so it would be good
if you file a feature request for this so we can be
sure to include it when we can.
>
> * Ability to view connections/references in the
> "catalog".
Do you mean as foreign keys under each table? Or as
a new subitem of each schema? Can you file a feature
request for that with more details?
>
> * Ability/option to hide connection captions.
>
That was implemented just recently, so it will be
on the next release.
> * Option to hide Indexes/Triggers dropdown if
> table does not have Indexes/Triggers (in default
> object notation)
>
Sounds sensible, please file a feature request too.
> * Collapse/Expand All:
> Tables/Views/Indexes/Triggers.
>
Noted.
> * Add a few default color options for labels
>
What labels?
> * Editing a table allows you to change foreign key
> cascading options, but editing a reference
> doesn't.
>
There's already a feature request for allowing
to access the FK options editor from the connection
one.
> * Views show up as nothing more than their name,
> ideally you want to see the end result of the
> view, e.g. a list of the final columns of the view
> and under a separator, a list of the tables the
> view uses.
>
That will require parsing of the SELECT statement,
so it's not a trivial task, so we'll keep that for a future
major version. Showing the contents of the query
in a tooltip is easy, however, so we can add that quickly.
>
> Bugs
> ------------------------
>
> * Right click, "Edit in new window" on a table
> does not seem to work any differently than "Edit
> Table".
This is fixed now.
>
> * In the OSS version, there seems to be no way to
> set the default storage engine for tables.
You can change the default engine in the top toolbar
for the table tool, but a global option for that
would be good, indeed. Please file a feature request.
>
> * Editing an advanced option in Tools->Options
> turns the option into a textbox, but when you hit
> enter after you've changed an option, it closes
> the entire window.
Please file a bug report.
>
> * Column definitions seem to go out of bounds of
> the table using the default font.
> table1_id |
> long_column_name VARCHAR(45)
> |^ table is cut off
> here, part of the 4 shows off the edge.
I think I've only seen that happening on zoom levels
different from 100%, can you confirm?
>
> * Object/relationship notation don't seem to get
> put into the undo history.
Please file this as well.
>
> * Large databases (with over 50 tables) are
> basically unusable in model view.
That's a known problem.
Thanks for the feedback! We won't be able to implement
everything you suggest immediately, but it's always good
to have an idea of what's urgent and wanted by users so
we can prioritize accordingly.
Cheers,
--
Alfredo