[re: signing weblog comments] There could be a niche for a minimalist “sign this text” application, if PGP were to soak into general net infrastructure. - Mark Pasc
Pondering the comment issue, I've come to the same conclusion. PGP and GPG are a little too excited about high levels of security, and seem to feel that it's more important that everybody immediately jump to 100% then to allow a more gradual use of the system. (As a result, nobody does; a better way to get people's toes in the water then WWW commenting I can't imagine, but that's not an option, so instead of a 90% solution, we get a 0% solution.) I'd like to see a key type that explicitly says "I'm a low security key! My public component may be hosted on a somewhat insecure webserver, and there is at least some responsibility on the part of my owner to make sure I'm still secure!"
It's good that PGP/GPG make it possible to have a really, ultra strong key tying into an ultra-strong web of trust that makes it almost impossible to get someone's private key, but I don't think it is good that it seems to require that mode of operations. There is a serious need for lower-stakes, lower difficulty, lower security signing. When signing comments, I don't really care that you know that I am THE ONE AND ONLY JEREMY PAUL BOWERS VALIDATION NUMBER #984375249875, I mostly just care that I can validate "I am the same person who said that comment over there, and here is the signature to prove it".
Incidentally, I'm planning a mini-review of PyDS against Radio Userland in a while here, but here's one thing I've gotten to work in PyDS I could never get in Radio Userland (without doing unsupported hacking on the Weblog core): Posts without titles, that actually can modify the generated HTML instead of just sticking an empty string in the template.
As you can see, I like to put boxes around my titles, and in Radio Userland, an empty title results in an empty box. Yucky. So I always wrote a title, even if I had to stretch. I'm thinking this may have inhibited some of my more off-the-cuff comments... like this one.
I suppose it will be for you to judge whether this is a good thing or a bad thing.
Reading the comments in this Roger Simon post about the President's Sunday interview reminds me of one of my fond wishes: Just once, I'd like to hear the President publically say something like "You know, it's easy to point out that I'm not the best speech giver in the world. It's a lot harder to plot the course of the free world." Something to just cheese off as many pundits as possible, while being politically a null statement (except for the effects of cheesed-off-based punditry). Along with being fun for me, it might even prove to his benefit since the Democrats really need to control their bretheren who feel inclined to put forth rants of questionable sanity; those folk would be unable to resist frothing at the mouth.
Probably not a good thing for discourse in the short term, but anything that cycles the ranters out of politics faster is fine by me. I largely agree with the President's Iraq policies but I am truly disturbed by the lack of credible opposition; even the good points are too frequently buried under intolerable levels of screech.
In my last post, I dissociated the concept of "outlineness" from a
graph, and showed at least the skeleton of a data structure that allows
the power of graphs while preserving the nature of an outline.
In this post, we will fix a flaw in the model built up to that
point, which is that there is no way to obtain a list of parents,
given a node, only a list of children. For various reasons, this is
necessary to building an outliner, so this flaw must be fixed.
Rand Anderson's (update: sorry about the misspelling!) previously-mentioned blog post brought his Executable Abstractions weblog to my attention, and I noticed something interesting about it; namely, that while it was obviously a generated weblog it didn't match the profiles of any of the many blogging systems I am already familiar with.
A little digging on the homepage led me to PyDS, which is basically Radio Userland in Python. (Not Frontier; the aspects that Radio Userland added to Frontier, the News Aggregator, the Weblog, Tools, nice Preferences screen, etc.)
Intrigued, I emerge'd it onto my Gentoo system. As I mentioned recently, the only full-time Windows XP machine left in the house is running Radio Userland, and I've been looking for a way off. I thought I'd eventually build one about a year from now based on Iron Lute, but it turns out that PyDS is more then capable of handling what I used Radio Userland for, so I decided to jump sooner rather then later.
(The benefits are that there are now two computers that used to be on 24/7, the Radio Userland computer and a computer that was providing secure access to it remotely; I'm almost afraid to find out how that was impacting my power bill...)
I've completed that transition. What you're reading is no longer Radio Userland generated, it is PyDS generated. I've also transitioned away from /irights to /iri; with the help of some mod_rewrite rules in Apache you should not notice any differences. (I've special-cased /irights/rss.xml to point directly at the correct RSS file without a redirect, in case your aggregator might be lost by the redirect.)
So far I am pleased; there's a few changes I want to make and submit patches for, but by and large it's a very, very solid 0.7.0 release. I'll be posting a review later on, bearing in mind that it's a 0.7.0 product and doesn't cost anything.
Please e-mail me if you experience any problems with this site that may be related to this switchover; that includes broken links (all links to old /irights URLs should still work) and everything else, except broken anchors (the anchors in the HTML have changed but that shouldn't affect very many people).
|<- Future Posts||Past Posts ->|