Forum OpenACS Development: I know that this is beta, but...

Posted by dwayne hinton on 10/09/01 04:39 PM

i think there should be a better way of loading modules for use. you see i'm really new to openacs, well not really new, i left and started using phpnuke for it's ease of install and use (i needed to get the site up fast, and ACS has too steep of a learning curve :)).

although phpnuke doesn't offer the functionallity, power, nor the security of ACS, however if you need to get a site up and running fast phpnuke wins hands down.

and that's my point, if openacs-4 could use some of the better functions of other opensourced CMS projects, like in phpnuke to load modules all you need to do is place it in the modules folder and the program automatically loads it. my reason for saying this is NOT to provoke your flames but to only point out that sometimes openacs can be difficult to use, and what good is software if nobody (well except, rich companies and the PHd's only they can afford to hire.) uses it. phpnuke and other CMS like it are new when compared to ACS, but have gained in popularity because of their ease of use, not because they r better then openacs. therefore maybe the lion(ACS) should learn from the mouse(newcomer CMS).

BTW, how do you load modules manually? the Package Manager gives an error. :)
just my two cents
dwayne

2: Response to I know that this is beta, but... (response to 1)

Posted by Jon Griffin on 10/09/01 06:17 PM

I would say the reality that PHPnuke gained popularity has nothing to do with its technical marvels but rather its use of the de facto Linux web scripting language (as crappy as it is).

3: Response to I know that this is beta, but... (response to 1)

Posted by dwayne hinton on 10/09/01 07:41 PM

i agree, but what i'm after is how can openACS become as simple to use as PHPnuke? i think the first step was the seperating of the tcl code and the cosmetics of the site. now how do openACS continue that process, like making it simple enough for the average web joe, like me :) to make use of it.
thanks for your response Jon.

4: Response to I know that this is beta, but... (response to 1)

Posted by David Kuczek on 10/09/01 07:53 PM

We have had a discussion on this topic some time ago. Members were talking about the improvement of the out-of-the-box installation...

https://openacs.org/forums/message-view?message_id=22628

Therefore you are right that this is still a problem of OpenACS.

Another thing that always interested me and that was brought up by Jon is: What are the pros/cons of PHP/Tcl...

I don't want to start a flame here, just interested in some true arguments that I can use for Tcl and against PHP...

5: Response to I know that this is beta, but... (response to 1)

Posted by dwayne hinton on 10/09/01 08:46 PM

i'm not sure out of the box is so good. will you lose some of the robustness of openACS? that's NOT what i (or probably anyone else) wants. well hats off and thanks to all you people whom have contributed to making openACS what it is, and my it continue to grow, cause i can't wait to include it in my tool kit :)

6: Response to I know that this is beta, but... (response to 1)

Posted by Don Baccus on 10/09/01 08:58 PM

I'm a little confused here. When I make a fresh install of OACS 4 I simply

Download acs-core and the packages I want from CVS someplace.
Configure AOLserver
Start up and let the bootstrap installer load the core
Reboot as instructed
Go to the APM and select "install packages"
Check off those packages I want to install
Check off more if I've checked off packages dependent on other packages (you might have to go back to CVS and pick up additional packages)
Let it install
Reboot AOLserver as instructed on the page.
Mount the packages using the site map.

At this point the site's functional. There are a bunch of steps involved but none of them are particularly difficult.

Now ... this isn't well-documented at the moment, but keep in mind that our first ALPHA (not BETA) release isn't until next week. We're going to be working on improved installation documentation in the next few weeks. Vinod Kurup has offered to step up and help Roberto Mello get this rolling ASAP. Good installation documentation will help a lot.

As far as an improved "out of the box experience" goes, one of the things we're discussing is offering some pre-configured package options, for instance as with the dotLRN initiative. We also will be providing RPMs that will help RH users, at least, get things up and running in a default configuration more quickly (AOLserver configuration, loading of acs-core, etc).

But such things will augment, not replace, the more general scheme which relies upon the APM and Site Map packages to load and mount packages. It would be nice to have a "mount wizard" that would make it easier to do both in one fell swoop, but the existing process, though a bit tedious, isn't difficult to learn - or won't be, once proper documentation is in place.

At some level, the separation between mounting and loading packages must be maintained, in order to support mounting of packages at multiple mount points (particularly useful with subsites). Subsite support does add some complexity, but also a lot of functionality...

7: Response to I know that this is beta, but... (response to 1)

Posted by Walter McGinnis on 10/09/01 09:11 PM

Don,

Don't forget configuring parameters and permissions after
mounting packages. In a lot of cases the defaults will work, but
not all. These are often overlooked steps.

Full documentation of setting up an OpenACS 4 site will be a
god send. Thanks in advance Roberto and Vinod.

8: Response to I know that this is beta, but... (response to 1)

Posted by dwayne hinton on 10/09/01 10:00 PM

i followed Jonathan Marsden's redhat 7.1 install to the letter. but when i go to the install packages link i get:
Please wait while the installer loads ........

HTTP/1.0 200 OK Content-Type: text/html; charset=iso-8859-1 MIME-Version: 1.0 Date: Tue, 09 Oct 2001 15:54:01 GMT Server:
AOLserver/3.3.1+ad13 Content-Length: 6027 Connection: keep-alive

Request Error

Entity: line 148: error: attributes construct error
type="query_file" path="www/presentation-acl-add-group-2.xql"/
^
Entity: line 148: error: error parsing attribute name
type="query_file" path="www/presentation-acl-add-group-2.xql"/
^
Entity: line 148: error: attributes construct error
type="query_file" path="www/presentation-acl-add-group-2.xql"/
^
Entity: line 148: error: xmlParseStartTag: problem parsing attributes
type="query_file" path="www/presentation-acl-add-group-2.xql"/
^
Entity: line 148: error: Couldn't find end of Start Tag
<file
type="query_file" path="
type="query_file" path="www/presentation-acl-add-group-2.xql"/
^
Entity: line 227: error: attributes construct error
<file type="query_file" path="www/slides-reorder-2.xql"/
^
Entity: line 227: error: error parsing attribute name
<file type="query_file" path="www/slides-reorder-2.xql"/
^
Entity: line 227: error: attributes construct error
<file type="query_file" path="www/slides-reorder-2.xql"/
^
Entity: line 227: error: xmlParseStartTag: problem parsing attributes
<file type="query_file" path="www/slides-reorder-2.xql"/
^
Entity: line 227: error: Couldn't find end of Start Tag
<file type="query_file" path="
<file type="query_file" path="www/slides-reorder-2.xql"/
^
Entity: line 230: error: attributes construct error
<file type="query_file" path="www/slides-reorder.xql"/
^
Entity: line 230: error: error parsing attribute name
<file type="query_file" path="www/slides-reorder.xql"/
^
Entity: line 230: error: attributes construct error
<file type="query_file" path="www/slides-reorder.xql"/
^
Entity: line 230: error: xmlParseStartTag: problem parsing attributes
<file type="query_file" path="www/slides-reorder.xql"/
^
Entity: line 230: error: Couldn't find end of Start Tag
<file type="query_file" path="
<file type="query_file" path="www/slides-reorder.xql"/
^
document was not well formed
error while parsing XML
while executing
"ns_xml parse [lindex $args 0"
(procedure "xml_parse" line 5)
invoked from within
"xml_parse $xml_data"
(procedure "apm_read_package_info_file" line 29)
invoked from within
"apm_read_package_info_file $spec_file"
("foreach" body line 2)
invoked from within
"foreach spec_file $all_spec_files {
array set version [apm_read_package_info_file $spec_file]
set version_name $version(name)
set package_..."

and more...........
if i miss spoke i'm sorry, or get this error because of my mistake please advise on how i could correct it thanks

9: Response to I know that this is beta, but... (response to 1)

Posted by Vinod Kurup on 10/09/01 10:12 PM

Looks like there are some typos in the wp-slim package's .info file. I've fixed them. If you're installing from CVS, just cvs update to get the fixed code.

Otherwise, go into the file packages/wp-slim/wp-slim.info

3 lines in this file are missing the closing bracket (>). Add the closing bracket to lines 148, 227 and 230. This should fix it.

Thanks for pointing this out.

10: Response to I know that this is beta, but... (response to 1)

Posted by dwayne hinton on 10/09/01 10:19 PM

big thanks!! that worked. there seems to be some version conflicts, but iseem to be on my way, thnx again

11: Response to I know that this is beta, but... (response to 1)

Posted by Don Baccus on 10/10/01 03:21 AM

Oops! Thanks, Vinod - I also fixed these problems on Monday but am in the midst of other changes and hadn't committed them yet.

Dwayne - the problem was a bad commit of a messed-up .info file by one of our porters combined with the APM"s overly-sensitive and easily offended personality that makes it intolerant of such errors. Over time I plan to put some time into making the APM more robust in the face of such errors - the package should be skipped but your entire install shouldn't grind to a halt.

Walter's absolutely correct, some packages at least need to be configured after they're mounted. I forgot that in my recounting above.

And ... Dwayne, thanks for your patience! The CVS tree will be broken from time-to-time, but overall people have done a remarkably good job of keeping it in a stable and consistent state. Remarkable considering that our team ranges from old fart software engineers like myself to undergraduate students in places like Cyprus and Australia, a non-hacker doctor who's become a monster contributor (go, Vinod!), etc etc etc.

12: Response to I know that this is beta, but... (response to 1)

Posted by dwayne hinton on 10/10/01 03:53 PM

YEAH!! we non-hackers can contribute!! :) that being said, i need to learn how to securely port the education module. can someone take a look and tell me if this is a do-able(i'm no en major:)) job for a newbie willing to learn? i mean i'll do the work but, i'm sort of good with php & mysql and new to tcl & postgresql. i also don't know anything about oracle (too rich 4 my blood). so let me know what you think. Thanks

13: Response to I know that this is beta, but... (response to 1)

Posted by Don Baccus on 10/10/01 06:31 PM

Sloan/MIT - the folks who sponsered the original education package (ACES) - are paying for a port of ACES to OpenACS 4. It will be called dotLRN - search for that word and you should be able to dig up some older threads about the project.

So in short, it's doable but it's being done. However, if you're interested in helping with other bits and pieces e-mail me. You could help with testing, you might be able to pitch in and help port code that's not been ported yet (though you face a learning curve since you're inexperienced with Tcl), you could help by either writing or reviewing documentation (for instance the install guide we've been mumbling about), etc etc etc.

We're not proud. We'll take your help whether you have 15 minutes a week free to contribute or 150 hours...

14: RPM-based install of OpenACS 4.x (response to 1)

Posted by Jonathan Marsden on 10/11/01 11:28 AM

Dwayne: Thanks for sticking with OpenACS4 in its current state!

If you have further specific suggestions for improving the OpenACS4 install from RPMs, do let me know. I've been rather inactive here for a few weeks, but hope to rectify that somewhat.

There even seems to be some code in the autoinstaller to try and mount some of the core packages automatically; it doesn't seem to work at the moment, though! I think it was a remnant from an ACS 4 classic autoinstall process that needs further work if it is to be useful to OpenACS 4.x. I do want to explore that further.

I'm behind my own personal schedule for creating OpenACS 4 RPMs... fortunately, OpenACS 4 itself has slipped too so my own slippage is not as noticeable!!

My main problem at the moment is trying to integrate the OpenFTS search stuff well into RPMs; it does things like reach back into its own source directory (eek!) at the moment, making a "binary" RPM install distinctly difficult. Getting Oracle search support working, and adding the capability to search more kinds of content, are obviously higher priorities for the OpenFTS team than improving the install issues right now, but I hope at some point to be able to solve this, and so create an easier "out of RPM experience" for newcomers to OpenACS.