(I owe a lot of the ideas in the following speculations to conversations with my wife Betsy, whose experience in the software industry mitigates and informs the pious hopery of academe... and I expect to revise this piece substantially)
The web offers an easy means to compose and distribute extramurally, to broaden the audience, and to extend the shelf life of a product, but most web pages are fairly 'flat' (don't have many links, and/or don't have many levels of links behind them) and most are also static: they deliver text, but usually don't offer much possibility for interaction. On the other hand, not much more than a word processor is needed to make such pages, and they can be very effective. Projects like this can be done on the fly, and are usually done by individuals, sometimes with coaching.
The next step in page complexity is often the addition of images, which entails the relatively modest expense of a scanner and image manipulation software, but quickly reaches the limit of what a non-specialist can do easily (especially once the features of complex and expensive software like PhotoShop become necessary).
Pages can certainly be made dynamic with dHTML, multimedia and Java inclusions, but the hardware and software to accomplish such prodigies is still more complicated and often needs priestly intercession (locally-produced CD-ROM courseware is another example of this level). This is where a media laboratory/training facility becomes a necessity, but the products are still for local use (and probably start to lean heavily on 'fair use' copyright interpretation). Products at this level are very likely to involve the collaboration of several people and take months to carry from conception to completion, and generally they are intended to have measurable curricular significance --but they're still essentially personal projects.
There's another level of product development to think about, involving a much greater commitment of time and resources, and aimed at a wider public than the face-to-face local group. This is software constructed to be distributed (and probably sold), and it's where the payoff for the hardware and software investment begins to be tangible --to be more than brave pioneering that makes the producer feel good. This is the level at which the facilities and services of a media center begin to be inadequate but where the product starts to have broader influence and significance --the media center may act as a bridge. We may not be intending to get there (in plans for media center and training lab), but I think we need to think about what goes on at this level and how we might get there.
A lot of today's courseware and educational software started out as local products, personal experiments, brave attempts. A successful local product may stay local, but it might take on another life in which it can really influence teaching in many settings --either via direct implementation or as a model that others build upon. There's a leap involved in producing a product that has the escape velocity to be useful in many places, and ultimately it means a commercial product with documentation and technical support. Freeware doesn't fit this model, which requires careful testing on various platforms and the specialist attentions of designers, software engineers and technical writers.
So a question for a lab setup is: what is success? Is it local (much effort for improved teaching in one course --the teacher's perspective), or does it have grander aspirations? Funding agencies probably don't have much interest in grandiose support for this level of development and purpose. Larger institutions have other priorities for their facilities, and it's worth looking at some examples to see what else we might do.
Northeastern University's Multimedia Lab was seen from the first as a development facility bridging the gap between local software and publishable authorship --or, to put it another way, between disseminating existing intellectual property and creating intellectual property. It provides a lot of services, can draw upon university-wide expertise to solve problems, and does a lot of putting people into contact with others who are doing kindred things, using the same software, hunting for publishers, considering copyright and trademark issues, etc.
The skills, energy, enthusiasm and technology to create intellectual property certainly exist at W&L. Should this be seen as a goal, as a part of pursuit of excellence in teaching? What is the vision here? At big universities, professors push the frontiers by research; we argue that we push the frontiers by excellence in teaching. So what are the frontiers of teaching? Facilitation of the ability of professors and students to innovate in excellence of teaching and learning is especially suited to W&L and similar liberal arts institutions. We should be prepared for the possibility that we might produce something astonishing, something that has import beyond the local, and we should think about what kinds of support that requires, keeping in mind that it's collaborative efforts that are necessary to change how teaching and learning can be done.