the Design Experience Weblog Archive

This version includes a plug-in API for authentication, locks, and properties. It includes better support for buggy WebDAV clients as well.

I really can't imagine a better way of decribing how learning really happens when someone is following something they are truly interested in. In addition to the support for the idea of learning coming from real interests is the long lost of problems inherent to schools such as stiled creativity, suppressed interestes, and curtailed freedom. If that is really what schools are like, why would anyone want to send their children there?

It is curious that in the context of increased use of technology such as weblogs, online communities, etc..., issues such as these are coming out. A networked society enables much greater interaction between people who are not in the same place and time. This is great for learning. One can contact an expert in any part of the world and learn. Of course it does not need to be an expert, just someone who knows what you want to know, and is willing to share. Even more interesting is a community of people learning together. So when I learn something new, I share it with my community, and hopefully from that, even more learning will happen.

These are the really exciting ideas. They are not new, but it seems the idea of learning anytime, anywhere, from anyone, is spreading with the increased connectivity afforded by technology. Of course, until everyone is given the chance to use technology in this way, the benefits will be small and scattered. How many kids are learning that they can learn much more outside of school being part of the real world because of this technology.

Folksonomies: How we can improve the tags from Lars Pind and The Obligatory Folksonomy Post from Michael Feldstein.

I need to dig into these and come up with something useful to say. Right now I'd say that they both mention looking at the resource first, and finding out how different people have tagged it. That is the opposite of finding what other resources have the same tags, and could produce another set of useful metadata.

* We released 9 versions of OpenACS, beginning with the first release of OpenACS 5.0, all the way to OpenACS 5.1.4

* Major changes and improvements to the toolkit, including:

o internationalization
o automated testing
o rapid development of tables (list-builder)
o external authentication
o support for Oracle 9i
o improvements to automated testing

* We've put in place (well, this is largely Joel's doing, I believe) a rigorous new release process for packages and for core. It took a while for it to start being used, but not packages are being released fairly frequently, and communities are starting to grow around particular packages.

* The installation process isn't much better than it used to be, but people are starting to work on it. The documentation is also very high quality, much better than it used to be.

* Registrations for the site are WAY up, better than they've been at any time since the inception of the project. I think we're finally starting to market ourselves better.

* We finally got the website onto OpenACS 5.x, thanks to Joel and Malte.

• Bug fixes.
• The missing CR TCL API has been filled in, thanks to Rocael and his team and Dave Bauer.
• This release does not include new translations.
• See the Release notes for a full list of changes.

Stephen says that since the top 10 results for a particular query don't change very often RSS is not useful for results from a search engine such as Google or MSN. I don't think you only have to look at the 10 ten results of course. There are probably some interesting items in the top 50 or so. I don't see any reason you can't ask for the second or third page of results as RSS just as well as the first.

This is probably more true if what you are searching for is covered by regular news media, or weblogs, or basically any web site that has regularly updated content. You can't dismiss search based on the fact that a basic query will return fairly static results. You can try an advanced query. If you restrict to pages that have changed in the last number of days or use other criteria to narrow the results I suspect RSS would be more effective.

Robin seems to focus more on the fact that somewhere in Stephen's post he mentioned new sites such as news.google.com. He mentions that there are many valuable sources besides news outlets, and this is true. He does also address the fact that simple queries will not return different results, and of course, you will need a more complex query to get valuable results.

This reminds me of the work I did for a client recently where the entire web site emitted RSS to share data with other affiliated web sites. The RSS feed generation was built on an advanced search concept. Full text indexes of content was one search criteria, but in addition, language, date, and a categorization system were also used to generate RSS feeds.

Why not return search results as RSS? Its as good as any other XML format. And who knows what interesting uses for the data can be imagined once its in a easily machine readable format.

One thing he says is that right now there are 600 photos at flickr tagged with the word "summer" and that after a while there will be 6000, or 60,000 and it will be difficult to pull out the ones that are relevant. Why just use the one word? What if you searched on "summer beach" or some combination of terms? This is how I use del.icio.us to manage links I want to remember.

I am not sure if I use the tags at del.icio.us in the same way as everyone else. I don't make up new tags by combining words, but just put in 3 or 4 words I think will help me find the items in the future. This is some sort of wacky variation of facted classification I think. Where I take out the most important aspects of a item and pick the word that defines that aspect. This means if you look at the list of tags I have, I have a huge number of unique tags, some with only 1 or 2 items in them. But if I search for an intersection of tags I can narrow it down quickly.

Anything that gets people to think about metadata is good. Maybe once you manually tag a bunch of items, the computer could examine those, and compare them to new items and suggest potential tags for the new items.