Monday, March 01, 2004 - Posts
The WS-Security standard
defines the nonce. It is used to prevent replay attacks. The nonce terminology has naturally transfered into the WSE.2.0 object model, where it confused me. Here's more from
Sam Ruby on this
elusive word and how to implement the concept.
Dear ATOM spec designers,
As a human aggregator of the blog-o-scape, I find a big hole in the current specifications/conventions for Weblogs:
Use Case:
- Feedster search for XOP.
- Find http://www.mnot.net/blog/ article
- Wonder who produces http://www.mnot.net/blog/
- Look for link explaining source of sight
- Find mnot.net at bottom of site
- Find "find out about me" link on homepage
However for most sites we have:
Use Case:
- Search for some topic.
- Find blog entry
- Go to root blog page
- Look for author/publisher information
- Don't find anything, not even a name
- Look at RSS feed to find author name
- Search for author name and hope to find a link that explains who the author is
Requirement:
The biography for publishers/entry authors should be codified in ATOM specs and supported by Blog site conventions.
Has there been any thoughts along these lines?
EWEEK is now
recommending IT managers to forgo improvements with current spam blocker tools and jump on board with several emerging technologies
fronted by Microsoft and others. Hopefully by year's end we can mark the end of spam. Woo-hoo!
Back in '91 when I started my career, I worked for a company that made an object-oriented development environment named Actor. It was a clunky-slow environment to work in although it had great features. About the time that we had memory and processor speed enough to make it a great environment, we helped to cook our own goose by developing ObjectWindows for Borland in Object Pascal and C++. In the painful ramp-up to master C++, I observed that we would be stuck with C++ until a new language came out with a new development platform and OS. Java was not that platform, because it tried to be all things to all people. Aim for the lowest common denominator and you often hit it. The CLI, in contrast is that environment.
Brad Abrams is thinking out loud whether we should be subject to a split personality for the CLR by re-introducing the ghastly debug build. However, reflecting on my earlier experience, I would say to let Moore's Law have it's way. The CLR is already quite fast enough for the vast majority of applications and in a few short years will be so for an even higher percentage. Let's not re-invent clunky solutions like the Real-Mode and shared multi-threading solutions of old that were overtaken by Moore's Law. Is that a world we really want to return to?
Update: Rico Mariani weighs in on Brad's suggestion.