Various opinions about BOINC:
1) Our top priority at this point is to build, test, and release clients and
wrappers
that work on the latest versions of all platforms, including Android, and with the
latest VBox.
Installing/using BOINC needs to be easy and seamless for everyone.
Currently this is not the case; e.g. the BOINC/VBox client
doesn't work on Win10 because it has an old VBox.
And there have been critical bug fixes since 7.6.33.
This process consists of many tasks: see
https://boinc.berkeley.edu/trac/wiki/AdminTasks
Rom Walton used to do most of them.
Since he left, most of the roles are unfilled.
We're trying to remedy this.
If you're interested in any of the roles, let me know.
2) The client and server software are in maintenance mode.
AFAIK they don't need any new capabilities.
They're already packed with features (like storage management) that no one is
using.
There are a few code areas that are overly complex:
app version selection in the scheduler comes to mind.
Other than these, the code is well organized, terms of both files and
directories.
And it's not clear that the cost/benefit of rewriting the problem areas is
worth it.
3) The GUI code is a mixed bag.
The BOINC Manager is sort of a mess code-wise, partly because of WxWidgets.
Also, the UI/UX needs to be rethought, with the goal of incentivizing non-techie
volunteers.
Eventually we'll need new GUIs for both desktop and mobile.
This is a long-term goal, and it needs to involve UI/UX expertise we don't currently
have.
4) The web code (PHP) is mostly in good shape.
The factorization into functions is nice; it makes it easy to change things;
e.g. using Bootstrap wasn't that hard,
and using a different CSS framework would be pretty easy if we wanted to.
A few areas (e.g. forums) have become overly complex;
I periodically try to simplify them.
The preferences code (esp. project-specific preferences) is a weak link.
We need to supply both simple and advanced versions, as the Drupal code does.
And we need to simplify the implementation.
5) The server documentation is a major weak link.
When scientists look at BOINC this is what they see - not the code.
We need to think about separating it more into "Getting Started" and
"Reference".
Right now it's just Reference, which is overwhelming to first-time readers.
-- David
_______________________________________________
boinc_dev mailing list
[email protected]
https://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.
