Forum .LRN Q&A: What are the .LRN packages?

Collapse
Posted by Joel Aufrecht on
Based on the CVS modules file, .LRN consists of, at least, two parts:
openacs-core
acs-admin
acs-api-browser
acs-authentication
acs-automated-testing
acs-bootstrap-installer
acs-content-repository
acs-core-docs
acs-kernel
acs-lang
acs-mail
acs-messaging
acs-reference
acs-service-contract
acs-subsite
acs-tcl
acs-templating
ref-timezones
search

dotlrn-all

acs-datetime
acs-developer-support
acs-events
acs-mail-lite
attachments
bulk-mail
calendar
faq
file-storage
forums
general-comments
news
notifications
user-preferences
survey
lars-blogger
weblogger-portlet
rss-support
trackback
workflow
curriculum
news-aggregator
news-aggregator-portlet
oacs-dav
xml-rpc
categories
dotlrn
dotlrn-admin
dotlrn-syllabus
new-portal
profile-provider
user-profile
bm-portlet
dotlrn-bm
calendar-portlet
dotlrn-calendar
dotlrn-portlet
dotlrn-dotlrn
faq-portlet
dotlrn-faq
forums-portlet
dotlrn-forums
fs-portlet
dotlrn-fs
news-portlet
dotlrn-news
static-portlet
dotlrn-static
dotlrn-homework
survey-portlet
dotlrn-survey
curriculum-portlet
dotlrn-curriculum
dotlrn-weblogger
dotlrn-news-aggregator
feed-parser

There is also an alias called dotlrn-extra, with:
evaluation
evaluation-portlet
dotlrn-evaluation
assessment
assessment-portlet
dotlrn-assessment
views
lors
lorsm
lorsm-portlet
dotlrn-lorsm
wp-slim
wps-portlet
dotlrn-wps
photo-album
photo-album-portlet
dotlrn-photo-album
random-photo-portlet

Are the extras part of the standard .LRN release? If not, should they be in the future? Should there be two different .LRN release, medium and large? If not, should people get these packages from the repository install?

Collapse
Posted by Rafael Calvo on
Hi
My personal view is that all packages should have a maturity level. All packages level 3 would be part of the standard dotlrn distribution and all packages level 4 that are also designed for eLearning would be the 'certified edition'.
If someone builds dotlrn applicatiosn for non educational purposes it does not have to be included iin the 'certified'
version

Does it make sense?

rafael

Collapse
Posted by Orzenil Silva Junior on
Thanks, Joel, for this work!

Just one correction: defined alias was written as dotlrn-extra but it is dotlrn-extras instead

As a suggestion maybe you could consider add lors-central (https://openacs.org/forums/message-view?message_id=294899) clipboard and related-items packages to dotlrn-extras. What do you think?

About dotlrn releases i think Rafael suggestion is interesting. But i think could be very difficult classifies dotlrn releases for educational and non educational purposes. As i see, dotlrn is ALL about education.

Maybe we could think about this distinction towards marketing places only and not as a vision for developers.

From a market point of view or for selling dotlrn we could have standard dotlrn distribution for e-Learning (that includes e-LANE code, for example), standard dotlrn distribution for knowledge management (that i hope in a near future could integrate code from COP software and dotFolio, for example) and standard dotlrn distribution for Personal Information Manager (that could result from integrating blogs, wiki, contacts and advanced search features into dotlrn).

As a vision for developers i think we could have just one dotlrn distribution and incite and support people in OpenACS community to share teaching and learning as a focus when developing not only for formal education as well as for situaded learning including communities of practice, corporate blogs etc.
Collapse
Posted by Joel Aufrecht on
We just discussed this in the weekly .LRN release meeting. There was agreement that the core should shrink, that things removed from core should instead be "certified," and that the total list of core + certified should not be smaller than what core is now. However, there was not agreement about how to certify, if "certified" was the correct label, and other details. Also, all changes will have to be after the 2.2 release.
Collapse
Posted by Malte Sussdorff on
A .LRN package becomes certified if the maturity level is at least three and at least three organisations using this module confirmed it's maturity level. The tracking of the confirmations will be handled by a new module called "installation tracker", which I will discuss and propose in a different thread (https://openacs.org/forums/message-view?message_id=318325).
Collapse
Posted by Nima Mazloumi on
I would recommend not to introduce a second term to describe the level of quality a packages has reached. I think that the APM maturity level is good enough. We simply need to discuss in the community what the measures are. Here some suggestions:

- number of prio 1 bugs
- number of prio 2 bugs´
- installation works?
- upgrade works?
- maintainer available?
- accessibility level reached
- automated testing worked well
- level of i18n reached
- unistall works?
- available for both oracle/postgresql?
- performance issues?
- does technical documentation exist (API, ...)
- does user documentation exist

There is a TIP discussing the maturity level here: https://openacs.org/forums/message-view?message%5fid=161393

Any comments?

Collapse
Posted by Malte Sussdorff on
Nima, I'm also against adding yet another measure, but... who does certify that the maturity level given in the info file actually matches the criteria. This is why I wanted to have a certified maturity level :).
Collapse
Posted by Dave Bauer on
These are good things to measure but without a maintainer, the rest are sort of meaningless. Number of bugs doesn't matter if noone is fixing them. Seriously a package can't be called mature if it has any bugs and there is noone responsible to fix, confirm, prioritize them. This is one of the major reasons for delayed release schedule for OpenACS and .LRN. The OpenACS core team takes responsibility for all OpenACS core packages in theory, and should delgate some of this to volunteers, if possible.

This leaves the huge number of optional packages. I think a volunteer maintainer needs to monitor the bugtracker and forums discussions about a particular package. It doesn't mean they need to fix every bug by themselves, but if the maintainer can't fix a bug for any reason, the maintainer should fix help, or another volunteer to fix the bug.

Another way to help clear up bugs more quickly is to have the submitter submit an automated test either directly to CVS or as a patch. This test should confirm the existance of the bug. This serves two purposes, one, it provides an easier way to reproduce the error, and two, provied an easier way to know when its fixed! Not every person who submits a bug has to do this, of course, but anyone with some knowledge of the toolkit can try.

So, whatever, rules you put in place to measure the maturity level of a package, you need someone to carry out and enforce those rules, otherwise the name will not have any meaning.

Collapse
Posted by Nima Mazloumi on
Dave, you are right but relying on maintainers is not practical. Most of the criteria I have listed are quantitative. So with every release you go through them and the level of maturity is calculated by that. No new release of a packages without this step taken by who ever is interested in a new release. If you want to enforce release don't indicate anything and the maturity level becomes very low automatically. The OCT can define a threshold a maturity level is not allowed to fall below. What do you think?

You can not ask for a maintainer. This is open source and folks commit themselves if they earn money at the end of the day (at least on the long run).