::fibreculture:: WebCT, Open Source and Beyond

[Chris Chesher]

On 09/08/2005, at 10:41 PM, Julian Knowles wrote:

> On 09/08/2005, at 6:07 PM, Chris Chesher wrote:
>> I've been using Moodle <http://moodle.org> for the first time this 
>> semester. After being lumbered with WebCT for so long, it's wonderful 
>> to start using a learning management system which has an information 
>> architecture and interaction design concepts in line with my 
>> pedagogy.
>
> Thanks for the link Chris. Can you point me to an 'unprotected' site? 
> It sounds interesting and I'd like to see it in operation on some 
> course material.

The moodle site is not 'protected'. It just requires that you go 
through a simple registration process before seeing some of the site 
(which is itself implemented as a moodle site). It's not a spam trap 
(but you might use a service like <http://www.mytrashmail.com/> if 
you're paranoid).


>> There are strategic reasons why public institutions should invest in 
>> FOSS. FOSS offers educational institutions a better business model 
>> than commercial software.
>
> When you say 'invest', what exactly do you mean? If it is open source, 
> the only 'investment' one can make is in the provision of human 
> resources for  development/programming. Are you suggesting that 
> universities hire programmers to work on FOSS development? Academics 
> who have programming skills are already contributing to a number of 
> FOSS projects either in their allocated research time, or via ARC type 
> projects. If you think about it, there might be several hundred kinds 
> of apps required to run a university... Would you see this as 
> realistic (ie hiring a small army of programmers?).
>
> It is also not clear whether you are advocating a switch to FOSS for 
> all tasks... or where you wish to draw lines in relation to this 
> 'business model'.

I'm not advocating open source for everything, or hiring new armies of 
programmers for the sake of it. In many cases there are good reasons to 
use commercial software, and no current working FOSS alternative.

Open source development practices could be applied more widely than 
they already are, particularly to the software needs that are specific 
to universities and other public institutions (and their 
constituencies). It is strategically important to support FOSS, because 
through collective will, they can become better and better:

1) Universities should more often investigate whether funds dedicated 
towards licences, support and customisation for proprietary software 
might better be diverted towards developing and customising open source 
alternatives, even where this is less attractive in the short term;

2) they should more often try to identify needs that are common across 
a number of institutions, and collaborate on building appropriate 
systems (the DSpace example that Katie brings up is an excellent case 
in point);

3) where emerging FOSS systems are potentially appropriate, 
Universities should strategically invest in bug-fixing, support, user 
testing, documentation etc. to bring them up to mission critical 
standards, or even look to outsource support services based on FOSS 
standards



>> At the moment, Universities not only invest large amounts to maintain 
>> software licences
>
> true, but if it works well, then this can be seen as a decent 
> investment, just like electron microscopes, or other specialised 
> resources.

>>  but spend even more supporting these proprietary systems.
>
> The way I see it , support costs are the same for commercial and 
> FOSS... Actually the latter is usually more expensive as the 
> documentation is usually poor/incomplete and the users are often less 
> able to help themselves via the RTFM method. Anyone who had to compile 
> and install Pure-Data (PD) before someone got it wrapped into a single 
> installer with a GUI will attest to that.

The costs of support may be comparable; but the difference is that this 
labour can potentially return value to the shared collective knowledge 
and code base of the open source community. Software developed in this 
way is likely to be less alienated from its sites of use, and 
encourages a participatory culture and better skills development in 
academic and support staff. It won't be upgraded simply because of a 
presure for returns from upgrades; nor will improvements be delayed or 
held back for commercial advantage.

The problems with documentation for open source software are changing. 
Open source software like the Plone CMS now has really good 
documentation, with several books dedicated to it. The open source 
development processes are also expanding to incorporate the work of 
people with wider expertise: such as writing documentation.


>> Support tasks are restricted by the inflexibility of black-boxed 
>> software (where open source can be customised or added to), and there 
>> is no return from this labour to the institution.
>
> example?

Compare WebCT and Moodle. One has standard off-the shelf functionality, 
where the other has a modular structure to which many different 
components have been contributed by different parties. The next of 
these components to be added is the Learning Activity Management Sytems 
(LAMS) developed at Macquarie University.

Some companies (including the developers of LAMS) are moving towards an 
open source model of software distribution, in which case they will 
earn their income from supporting their software.

The argument that students get cracked software anyway doesn't read 
well as an official policy position, even if it's always been true.


> Many commercial apps have plug-in architectures with publicly 
> available standards, so you can write your own plug-ins.. Furthermore, 
> some commercial apps are 'toolkit' based and specifically designed for 
> tool/instrument building (eg Max/MSP). I don't think the commercial = 
> closed, FOSS = open comparison always holds true. Some FOSS can be 
> quite closed, in that support for common standards can be poor to 
> non-existent.

Commercial software development is always subject to competing 
pressures between the attraction of controlling monopolising standards 
and the more dispersed value of being interoperable. Open source and 
open standards tend go together much more naturally.


>> When support staff learn to use and customise these systems, the 
>> institution becomes even more closely tied to companies which won't 
>> necessarily be there next year anyway, and don't provide 
>> comprehensive support.
>
> Which is why most large institutions avoid FOSS for mission critical 
> tasks.... There is no support or service guarantee. They are not going 
> to run their financial systems on FOSS if , when it goes down, they 
> can't have it brought back up by tech support in a very short 
> timeframe. It is often these service/support needs that drive choices. 
> For what its worth, my university's e-learning unit evaluated free  
> open source alternatives to WebCT recently, and decided to stay with 
> WebCT, not because they thought it was necessarily superior, but 
> because they could enter a very firm service/support agreement. 
> Sometimes, the decisions are made according to the very real demands 
> of running a very large operation with minimal tolerance of system 
> downtime.

That's where leadership for such initiatives probably needs to come 
from a higher level. And a lower level.


>> By devoting the same resources to support and develop open source 
>> infrastructures, the quality issues for FOSS will quickly disappear 
>> (and where they don't, as Anna points out this doesn't necessarily 
>> preclude learning)
>
> Although an attractive proposition, I'd like to see the empirical 
> evidence for this assertion. A massive development effort would be 
> required. To get some perspective, one software revision of a 
> commercial audio app I use for teaching included a million new lines 
> of code. If we multiply that by several hundred software 
> applications.... then...

Quality is not only an issue for FOSS. In audio production, one of the 
leading products, Pro Tools, was notoriously unstable and difficult to 
work with in teaching lab environments, largely because of its 
anti-copying strategies, and in spite of a short-lived experiment with 
a free version. This situation continued for so long because only one 
party could deliver improvements.

There are too many cases in other contexts where reputable commercial 
software is just not that great. If some of them were electron 
microscopes, you wouldn't make out anything smaller than a fist. 
Applications used in universities such as WebCT, Interwoven web content 
management, Peoplesoft or Hyperion are expensive to licence, but in 
many cases have been costly or problematic to implement, and are not 
well liked by many end users. They tend to force the whole organisation 
to fall into line with their conventions (e.g. changing terminology, 
workflows, etc). These companies have excellent sales and legal teams, 
are highly profitable, and offer attractive service agreements, but 
that doesn't necessarily mean that their software is that great. It can 
take enormous resources to get such systems running well, and yet all 
this work in customisation creates no independent or shareable 
intellectual property.

Across the world, Universities (and other public institutions) already 
have huge labour forces whose efforts might be better capitalised upon 
if they had the resources and opportunity to help develop, adapt and 
improve the software infrastructure that they themselves use. It won't 
happen all at once, but it could provide cheaper and better long term 
solutions: one case at a time. I agree that this vision can be clouded 
by open source ideologues. There are many problems with current open 
source options. But FOSS needs friends. It is constantly FUDDed by 
vested interests, and threatened by legal moves such as software ideas 
patents.


> Let's not forget that the FOSS community is already significantly 
> populated by university researchers (and their postgrads) To put it 
> bluntly, salaried academics are among the few who can devote so much 
> time to developing stuff which is given away for free.  Call me a 
> cynic but.... quite seriously...  I think this is a significant 
> factor. Seen in this context, universities are already making a major 
> contribution to the effort.

Exactly. It's already happening. It just needs to be embraced more: 
with some strategic co-ordination, better follow-through on promising 
projects, official recognition and legitimacy.



Chris




--
Dr Chris Chesher
Director, Arts Informatics
School of English, Art History, Film and Media
Room S314 John Woolley Building A20
University of Sydney NSW 2006 Australia

Phone: +61 2 9036 6173 Fax: +61 2 9351 2434
e-mail: chris.chesher at arts.usyd.edu.au
Web: http://www.arts.usyd.edu.au/departs/informatics/