Unblocking The Main Thread Solving ANRs and Frozen Frames
Will the real spatial web please emerge
1. Will the Real
Spatial Web
Please Emerge???
[or why we have to take matters into our own hands]
2002-12-19 adoyle@intl-interfaces.com
2. The “big” idea
• You can’t build a Spatial Web if you’re going to
IGNORE THE WEB
• Google & friends
• REST
• Blogs
• Wikis
• Usenet
• RSS
• Next Steps…
2002-12-19 adoyle@intl-interfaces.com
3. Google vs. Catalogs
• Catalogs are fine but they tend to hide
behind obscure interfaces (even if they are
publicly defined)
• Google - billions & billions of entries
• Entire NASA metadata holdings for
EOSDIS ~ 10GB.
bucket
– A drop in Google’s
2002-12-19 adoyle@intl-interfaces.com
4. REST
• Give me a link and I’ll crawl the world, give
me a CGI script and I’m stopped in my
tracks
• Meet me 1/2 way… How about dual-entry
catalogs? Browsers, bots, spiders, &
crawlers. (Did I mention Google?)
Z39.50
OGC Catalog
UDDI
Query Engine “Link Portal”
…
Catalogs
2002-12-19 adoyle@intl-interfaces.com
5. Blogs
• Web Logs
• “Latest on top” series of semi-random
articles
• Comments allowed but usually a level
deeper
• Structure imposed by Blog package
• RSS describes top level
Content can be found in (yup!) Google…
2002-12-19 adoyle@intl-interfaces.com
6. Wikis
• WikiWikiWeb - a communal experience
• Free form, internally interlinked
• Structure imposed by content authors
• RSS about updates to a page (not always
useful)
Google… it’s all there, just leave a link somewhere.
2002-12-19 adoyle@intl-interfaces.com
7. PURLs
• Persistent URLs
• Either make sure yours never change or use a
service. (Former is more scalable).
• Use PURLs instead of UUIDs
• ONLY USE RESOLVABLE URI
– Even if you insist on URNs or outsource your PURLs
Will Google crawl URNs? How about PURLs? Hmmm…
2002-12-19 adoyle@intl-interfaces.com
8. What makes the Web GREAT?
• Google?
• Well, yes.
• LINKS?
• Well, YES!
• If you are a content provider, you should
provide links to your content!
2002-12-19 adoyle@intl-interfaces.com
9. Usenet
• All Usenet postings since ca. 1980 are
available on… Google.
• We have alt.<your favorite kind of deviant
behavior here>, searchable on Google.
• Why not alt.metadata?
• Why not alt.metadata.opinions?
2002-12-19 adoyle@intl-interfaces.com
10. RSS
• Really Simple Syndication or RDF Site
Summary
• Structured information about content
• Blogs & Wikis generate it
• Can be generated by other programs as well
• Example (won’t work for you…)
http://127.0.0.1/mywiki/RecentChanges?action=rss_rc
2002-12-19 adoyle@intl-interfaces.com
11. Next steps
• Let’s go flatten a catalog
• Let Google find it
• Work on a string-based location scheme
• Develop some useful tags
• Build some Blogs, Wikis, etc.
• Flood some poor newsgroup with metadata
• Generate some data discussion
• Play for 2-4 months. Google will get it. Then we
can decide if it was worth it.
2002-12-19 adoyle@intl-interfaces.com