The OpenACS Directory
Created by Gustaf Neumann, last modified by Benjamin Brink 29 Jun 2017, at 04:35 AM
Created by Gustaf Neumann, last modified by Benjamin Brink 29 Jun 2017, at 04:35 AM
Created by Dave Bauer, last modified by Benjamin Brink 29 Jun 2017, at 04:34 AM
See Also: openacs-release-status openacs-todo
Next Meeting: 2010-04-28 18:00 CET/CEST Convert to your local time
Created by Rocael Hernández Rizzardini, last modified by Benjamin Brink 29 Jun 2017, at 04:32 AM
Agenda:
1. Zen vs. acs-core
1.1. Site master template will have to be changed to be compatible with package changes for zen
In particular content pages need to be wrapped in <div class="main"> (or id= or something)
Solution: we do <div class="main"><slave></div>
1.2. CSS
Don propose we duplicate CSS at the moment rather than try to integrate .lrn zen theme css and core - our schedule's tight as tight can be and this would be faster
Someone has to do it (next meeting we'll determine who does what and probably the overall work plan)
Change to be done in oacs-5-3.
Sharing a single css file comes for 5.4
2. oacs 5.3.1 release
2.1. The zen work above
2.2. Greek translations
2.3. Nothing more other than critical bug fixes that hopefully are well-tested
2.4 Target date: 1st. march alpha.
2.5. NOTICE: people who upgrade existing sites will have to change their layout! To what extent is something we have to figure out in the next OCT meeting.
2.6. To work on the layout enhancements:
check this page:
signup for something
contact: honchos@dotlrn.org which is the people leading the effort.
.lrn mtg every tuesday at 1700 GMT
2.7. We don't have time to deal with UI clean-up for 5.3.1
Present: lee, don, dave, roc.
Created by openacs irc community, last modified by Benjamin Brink 29 Jun 2017, at 04:31 AM
This is from a discussion about permissions on OpenACS' irc (names changed) that took place sometime circa 2005.
ryan: How do I create a group containing other groups? dave: composition_rel ryan: For instance, I have 30 admin groups, and instead of adding user A to each one manually, I want to add her to one group, and thus all 30. dave: what are you trying to accomplish? dave: you can't do that ryan: crap. dave: it it totally non-intuitive dave: here is why :) dave: we have the Super Admin group dave: wait. dave: no it doesn't work ryan: So what is a composition_rel? I thought parties were supposed to be a super-set of groups. dave: let me explain :) dave: no ryan: ok, thanks :) dave: here is how it works. dave: Super Admin dave: then we have admins_a which is a component of super admins dave: maybe it can work. dave: question is dave: can a group dave: have a composition_rel to more than one other group ryan: So what is the definition of a component? dave: lets find out. dave: a component dave: so if Admin A is a component of Super Admin dave: then every member of A is a member of Super dave: which is NOT what you want. dave: you want ever member of super to have permission over all the groups "inside" it right? dave: but in this case every member of A would have permission over all the other components etc. dave: group are NOT for permissions. dave: that is the design weirdness ryan: huh? ryan: Now I am completely confused. dave: you can't use groups the way you want ryan: isn't the whole point of groups to avoid permissioning on individual users? jim: but you should be able to build a page that asks for a user and puts the user into the 30 groups dave: ryan, yes. dave: you they don't inherit the way you think ryan: so you set permissions on a group with a set of objects, then just add/remove users from the group, right? dave: its backwards to what you are thinking. dave: yes dave: that works dave: perfectly jim: so you can get what you want (convenience, non-tedium) but have to do it another way dave: but composition_rels dave: behave backwards dave: they are not useful for org chart models ryan: ah ok. dave: but I think it can work ryan: What is an application of composition_rels? dave: here is what you would do dave: if it works dave: create all your groups dave: Admin A, Admin B etc ryan: done. dave: then make one group dave: and give it a composition rel to all of those groups. dave: its upside down. dave: then if I am in the one group, i am in all of those other groups jim: so you're putting the one group into all the groups jim: that should work :) dave: yes d ave: b/c its not _in_ dave: its a component. d ave: i think that will trigger the correct permissions. ryan: how do components work in the data model - I want to understand this better. dave: well jim: it' dave: then I recommend you 1) read the acs-kenrel sql files jim: s a special kind of acs_rel dave: 2) run alot fo experiments in psql dave: 3) find a bug in the triggers dave: :) dave: that is how I figured it out. dave: sucks huh. dave: seriously the comments in the SQL files in acs-kernel are very illuminating. dave: also have you read permissions tediously explained? ryan: but you're it'll work? ryan: Yes. d ave: i am not sure it'll work dave: but jim: we think it will work r yan: but I could re-read it a fourth time dave: i don't see any rule dave: that disallows a group being a component of more than one group dave: if there is such a rule, it won't work. ryan: but compositions typically extend 'up' the chain of groups? dave: yes dave: that is what its for dave: so for example ryan: what is a practical example? dave: I have Main Subsite dave: and several other subsites dave: wait dave: actually this is an example of why it doesn't work :) dave: hmmm actually I have to check dave: not sure if susbsite groups are components of main subsite or not. dave: .... ryan: You see, I want to create this super group and then let the client admin the members... dave: ok there are no rel_constraints on a default install. so that should be safe. dave: yes d ave: but its really a sub-group jim: try it with two groups and another group be a component of both groups... give each of the first two groups two different permissions... put a user into the component group... dave: a super group would not work the way you want. dave: here i what you do jim: see if the user has both perms dave: 1) create two groups dave: 2) create another group dave: 3) make the third group a component of the first two dave: 4) add someone to the third group dave: 5) check if they are a member of 1 and 2 dave: for extra credit dave: apply a permission to 1 and 2 dave: check if the members of 3 have permission on those things dave: if this works jim: 1-4 and the extra credit are what I just sed :) dave: i just solved the oldest OpenACS 4 riddle joe: As a topical aside, we changed the way we use groups in our subsites for dotcommunity. One for admins and one for members (so we don't use admin_rels). Then we use a composition rel to make admins of top level sites also admins of lower level sites, and to compositions in the opposite direction to make members of lower level sites members of the higher level ones too. dave: jim, yes :) jim: 5 is a good idea too ryan: ok, sounds good. Will test and get back to you. That would be very cool if it could work both ways. Obviously it is important to be able to have groups of groups... dave: joie: you could still use admin_rels to do that dave: it would work the same way. dave: ryan, yes if that works the way I expect it would be cool. dave: joie, so is the lower level admin group a component of the higher level admin group? joe: The problem with admin rels was that an admin of a subsite became an admin of the supersite, which isnt what we wanted. dave: or the other way around? dave: joie: yes dave: that is what I just said dave: the component rels go the wrong way dave: than what you would think intuitively dave: although mathematically they work correctly as specified. dave: damn PhDs dave: basically we need to write high-level functional wrappers over all this crap joe: So we have a composition rel going "down" for admins, and "up" for members. So the supersite admin group is a component of the subsite's admin group, and the subsite's members group is a component of the supersites members group. dave: so you can just call a tcl proc that tells you what happens (instead of what is does in the database) dave: joie: ah so it _does_ work. that is just what I told ryan-g to do dave: we need to write a tcl api to do that that is clear what is happening. jim: so members of the subsite become also members of the supersite dave: yes dave: which makes sense joe: Indeed. Then we frigged the acs-subsite members pages to do the "right thing". dave: but then admins of the subsite become admins of the supersite (if you use admin_rels) dave: which is why you don't want to do that. dave: joie: but you are right dave: and openacs is wrong dave: except I doubt there is an upgrade script that would work dave: damn PhDs dave: joie: why the hell didn't you tell us this before? :) dave: i have been trying to figure that out for 4 years joe: We have an upgrade script that does it somewhere. Rob wrote it. dave: you rock. joe: We weren't sure the new way was "right". dave: yeah dave: it is dave: b/c it makes sense dave: well dave: except dave: no its right. joe: We then got rid of admin_rels completely. joe: and only use membership_rels dave: b/c everyone expects it to work that way dave: yeah see dave: the problems is dave: all this stuff was experimental dave: and no one every finished it jim: you would need to be careful when deleting certain objects dave: except you did. dave: so now we can say "this is the way its supposed to work' We can say that because that is the way every one has expected it to work, but it never did dave: wow jim: make sure to remove all rels first then delete dave: i am so surprised. joe: The code is in the dotCommiunity download on www.dotcommunity.org. The upgrade scripts aren't there though. dave: this is so cool. dave: get them! dave: :) joe: Heh. I'll talk to Rob about it tomorrow, as not sure what he implemented.
Created by Robert Taylor, last modified by Benjamin Brink 29 Jun 2017, at 04:28 AM
See: OpenACS News Section
Created by Dave Bauer, last modified by Benjamin Brink 29 Jun 2017, at 04:20 AM
OpenACS (Open Architecture Community System) is a toolkit for building scalable, community-oriented web applications. OpenACS is the foundation for many products and websites, including the .LRN e-learning platform. OpenACS is open source and is available under the GNU General Public License.
OpenACS is unique in the breadth of services it offers developers and administrators. Millions of dollars and decades of developer time have gone into the maturation of OpenACS.
You can read the technical reasons to use OpenACS.
The installation documentation contains all the necessary steps to install OpenACS on a large variety of systems. There is also a Windows Installer and other packaged installations openacs-system-install. Also Check the installation requirements before installing. The current stable release is OpenACS 5.9.0 (cvs aliases).
You can start by reading the documentation, specifically tips on customizing, the developer's tutorial, and
the FAQs. There is a OpenACS Wiki and a list of packages that extend OpenACS.
For professional help, contact one of the OpenACS companies.
One of the strengths of the OpenACS project is the community surrounding it:
irc://irc.freenode.net/#openacs
3445 files changed, 215464 insertions(+), 193642 deletions(-)
The release of OpenACS 5.10.0 contains the 93 packages of the oacs-5-10 branch (plus 9 compared to oacs-5-9). These packages include the OpenACS core packages, the major application packages (e.g. most the ones used on OpenACS.org), and DotLRN 2.10.0. The release contains several security improvements and new functions, several new packages were added, and 370 bugs noted in the bug tracker have been fixed (plus many more not noted there).
For a partial summary of changes, please check the release notes [2], for the more detailed list of changes since the release of OpenACS 5.9.1, see [3]. Many changes/enhancements of the application packages are just contained in the detailed changelog.
These changes were contributed by 7 committers ...
... and additional 13 patch/bugfix providers
All packages of the release were tested with PostgreSQL 13.* and Tcl 8.6.*.
[1] https://openacs.org/projects/openacs/download/
[2] https://openacs.org/doc/release-notes
[3] https://openacs.org/changelogs/ChangeLog-5.10.0
Created by Robert Taylor, last modified by Benjamin Brink 29 Jun 2017, at 04:18 AM
This page remains for legacy discussion.
This is regarding Approach 1. in en:Documentation_Project_Discussion. The milestones are performance based, instead of directly connected to work, which makes them an impractical plan to follow.
Here are a few things we are working to accomplish with XoWiki and the OpenACS website:
MILESTONE I: Tooling Up - Before We Can Accomplish Anything We Must Have a Powerfull Wiki.
- XoWiki is functional for the most part
- a few bugs and feature requests remain, Gustaf and Dave have rocked the OACS boat and created some amazing functionality in a short period of time. When some of these items are cleared this Milestone will be considered accomplished.
MILESTONE II: Testing Packages after each OACS release.
- XoWiki is functional, we can start testing immediately. Look for a TESTING PROCESS document in the PACKAGES category (if it's not up now, soon).
MILESTONE III: Look and Feel vs. Content
- XoWiki is turning out to be very powerfull and easy to use. We will need to focus on content first, look and feel can be handled later. UTILITY has a greater priority over LICKABILITY.
MILESTONE IV: Fanatical Documenation.
- API documenation will stay as is.
- see en:Documentation_Project
MILESTONE V: New Blood.
- I think we can start attracting new devs as soon as our Documentation allows a new developer easy egress into the toolkit. Watching Dave, Gustaf and others litterally create new functionality out of thin air over the last week of January 2K6 is clear indication the barrier to entry is not the toolkit, it is with the resources and tools they need to be able to understand and become developers.
- I think getting a LICKABLE look and feel is something that can be done simultaneously, at a stage after XoWiki has been thoroughly debugged and tested.
Created by Olga C. Santos, last modified by Benjamin Brink 29 Jun 2017, at 04:11 AM
1. Analyze the accessibility status of .LRN/OpenACS and discuss the existing main problems and the way to solve them.
2. On-going developments and research works on accessibility compliance in LMS
3. Educate .LRN community in the importance of developing with accessibility requirements in mind
4. Define some guidelines to help .LRN developments reach the accessibility requirements
Martyn Cooper - Head of Accessible Educational Media team at the Open University in the UK
Among diverse research and internal consultancy roles Martyn Cooper has overall responsibility for accessibility in the Open University's VLE which is based on Moodle. The Open University has nearly 10,000 disabled students and takes its legal and moral responsibility to give them equal access to its teaching and learning very seriously. It has been making substantial investment within the Moodle community to address the current deficits in accessibility of the software. This paper reflects on this process and more general issues of accessibility in community based and open source software developments.
You can send the proposals for papers to the Workshop call or to the contact below.
For questions contact Olga C. Santos
R&D Technical Manager
aDeNu Research Group (UNED)
Last modified: 2017-06-29 04:11:17.78443+02
Created by Matthew Burke, last modified by Benjamin Brink 29 Jun 2017, at 04:09 AM
Listing of potential projects for the Google Summer of Code 2007.
.LRN is serving as the mentoring organization for this project, but because .LRN is almost exclusively built on top of the OpenACS Web Application Toolkit project ideas that benefit one or both are welcome.
Suggestions:
Created by Dave Bauer, last modified by Benjamin Brink 29 Jun 2017, at 04:07 AM
Here is a list of topics people have expressed an interest in discussing at the LRN/OpenACS Fall Conference 2006
For those interested in participating in the 'what is the potential of a kernel rewrite?' discussion, here is a set of questions to think about. They are aimed at establishing the general point of view of those taking part in the discussion; ie. you don't need to answer them now or post a long article justifying your perspective. Simply knowing your answers and being able to compare them directly to other people in the discussion will help us get something useful out of it. You may also find it useful to consider these questions first from an optimistic perspective and separately with your own assessment of what is realistic (as people will have wildly different opinions of what is practical which may in turn hide the fact that people agree in principle).
What state is OpenACS in now?
Would you like to see all effort focused on incremental improvement? radical change? both in parallel (if so how should it be organized)?
What is OpenACS?
If your not sure what it is now, what do you think it should be?
Should OpenACS require local customization or be entirely configurable?
What are your aims for an improved OpenACS?
How should the OpenACS architecture be decided?
What do you like about OpenACS now?
What are the biggest problems with OpenACS now?
What is your OpenACS wishlist?
What are your OpenACS implementation plans in practice?