In the released version it is possible to filter by max 300 projects. Now 300 is not a small number to apply a filter, but if you have thousands of projects in a single tree, it becomes difficult and long to restrict your filter by only a few branches. Add to this the fact that each time that you select a project a textual preview of the project is requested to the server and what you get is a serious usability problem.
In the beta published today there are a few changes that simplify filtering by a large number of projects:
- Automatic textual preview of project is disabled in the Projects Tree. You can still view the record from the context menu. This allows you to click and select/deselect projects quickly
- You can specify the option to check automatically the children (down to the n-th level) when you check a parent. This feels more natural: if you check a project you probably want to display also the documents in the sub-projects. If this is not the case you can quickly deselect the sub-projects that should be excluded.
To enable this option add the following line to the file in DBWorks\WebViewerOptions.txt:
select_subprojects_automatically - The limit of 300 projects has been removed and made parametric.
To enable this option and specify that you want to show a warning for 1000 projects add the following line to the file in DBWorks\WebViewerOptions.txt:
max_projects_filter 1000
Note that if a user applies a filter by a higher number of projects he can still apply the filter but he is warned that filtering by a large number of projects can generate complex queries that require longer to execute and overload SQL Server. We suggest that you perform a few tests and configure the parameter so that the users are warned only when/if SQL Server slows down under the strain.
The value you should indicate depends a lot on the Server hosting the SQL Server database, especially the RAM available.
As a rule of thumb 1000 projects can be handled correctly, but it’s a trade off between usability (number of clicks) and performance. In these cases nothing can replace good old empirical tests.
Filed under: Beta, feature, Performance, Usability