I like this discussion. Eric Scoles wrote: > > > On 2009-01-12, *Dana Paxson* <[email protected] > <mailto:[email protected]>> wrote: > > > True -- the Web gained immense leverage from its ease of > access. But so > did a lesser example: CB radio. They created a commons, in each > case: a > place shared for general good. > > And then began the rise of the hooligans -- in both cases. The > commons > got raided and seized by those who only wanted to own them and exploit > others. CB radio became a useless fog of noise. The Web has become a > home for professional malware and financial exploitation. > > > > But unlike CB radio, our economy would utterly collapse without the > web in its current, stateless (one-way) form. And in any case, you > might just as well have described early (or modern) cities: Homes for > professional malfeasance and financial exploitation. Oh yes. Examples abound. I used CB radio only because it was vivid for me, not because its scale matched that of the Web. You cities example is better, but cities were often formed for reasons other than common good, unless one considers the good of the nobility to be the common good. (But then, the Web has its own über-geek nobility in there somewhere... (;-) > > > Our hindsight > tells us that two-way links might have helped -- they might still > help, > if we can find the geniuses who can implement them cheaply and usably. > > > > Implementation cost is no longer the issue, as it was with Xanadu (and > OpenDoc, my other analogy). You yourself could invent a half a dozen > more or less workable methods of implementing 2-way hypertext across > HTTP with an hour's thought -- I've no doubt of that. You needn't even > change HTTP -- you could use a layered approach to add meta-protocols > onto HTTP. What is a protocol, anyway, but an agreement to communicate > in a certain way? Just add new layers to that agreement. HTTP is > dirt-simple; the result could be dirt-simple, too. > > The critical issue now is critical mass of acceptance: new > protocols would not useful unless enough interesting pages used them > that people started to regard them as useful. (Or that systems started > to regard them as useful. People wouldn't need to be aware of their > existence.) One of the benefits of using HTTP is that you need not > even have the same 2-way protocol operating on all pages, as long as > there was some means of inter-operating. A page could load a script, > for example, that maintained a DOM object mapping Dana's 2-way > protocol with, say, Tim B-L's. Or it may even be sufficient to simply > add extra arguments to the same anchor tag: One for the Dana protocol, > one for the TBL. Who knows -- I haven't heard of anyone trying it yet, > but I'd be amazed if that was due to anything aside from my own > isolation: People MUST be trying this kind of thing. > > (Aside: It seems to me that Jabber is likely doing something like a > 2-way link, already, since it's essentially HTTP driven and > session-bound.) Thank you -- these observations open the picture. I like your interop suggestions, regardless of whose name is attached, and I'm sure you're right: it's being done as we speak. The real problem is, as you indicate, acceptance. > > > After all, we've built spam filters and anti-malware code of great > sophistication and power; couldn't that same ingenuity be applied to > redesigning TCP/IP and fitting it forward to prevent what we > suffer now? > > One-way links have the terrible weakness of anonymity of their > source -- > a virtue in a trustworthy setting, but a weakness in the general > world. > > > > A weakness, but also a strength, because it allows me to be > unconcerned with getting the approval of anyone I link to. I regard > that as a great boon to free society, by way of encouraging the free > flow of information. Two-way links imply two-way approval: > I would have the right to deny someone the ability to link to me. > Personally, I think that's wrong, and I think it would be liable to > kill network viability. Well, doggone it, I've had a bellyful of one-way trash flung in my face, and I am not receptive to just any old nutjob waving text and images and oh yes, malware at me. Just this morning I got a spoof e-mail telling me someone else had charged several hundred dollars to my PayPal account, and I had to log in, check my transactions, and report the spoof to PayPal. Total ****ing waste of time. I get this crap every day. It makes up a big chunk of my e-mail, and burns up a big chunk of my life. Check spam filter #1, pass on any legit e-mails that got stuck there, check spam filter #2, ditto the legit e-mails stuck there, check spam that got through #1 and #2, and so on. When you've got a bunch of e-mail addresses serving different purposes, multiply the total time wasted. Arrgh!
Network viability? Exactly what percentage of network bandwidth is eaten up with pure garbage? A lot. Two-way links would be a nice way of stopping the crap altogether, and then we have more netwidth for... well... watching sports and movies. Hmm. Maybe I'm doing wrong thinking. (8-D But network viability might be improved if we knew who was calling. It works very nicely for phones -- the telemarketers are a lot easier to spot now than they were before Caller-ID. If I want to send a message to someone who doesn't have my e-mail address on an approved list, I have options: go to Facebook, go to LinkedIn, go to SL, go to MySpace, e-mail a mutual friend, get social, do my networking, and form the connection I need. The sheer number and variety of social networks online is making such an approach increasingly easy. > > There's an analogy that's liable to be drawn between HTTP linking and > person to person networking. Metcalf's law applies profoundly to the > latter, but not so profoundly ot the former, due in large part to the > lack of 2-way authority on links. That's as it should be. Metcalf's > law has been taken far too seriously for far too long, anyway. > > > E-mail has similar weaknesses, also designing in the anonymity of its > source, and that took us into the tornado of spam and malware. Now > we've at least set up filters, but they're a major pain -- we're > addicted to convenience. > > > > "Addicted" is an interesting word, and certainly valid from one > perspective. But another way to put it -- not contradictory -- would > be to say that our economy is reliant on that arrangement. As with > HTTP, though not as profoundly: We could restructure SMTP to remove > many of the vulnerabilities and effect a swtich that took in most > economically-active users without losing the essential power of email. > It would cost, though. I think the real argument has been around who > gets to reap the cost. > Always true. But the patience of many is wearing very, very thin. At some point the failure of patience will drive a solution. It may not be a good one -- those last-ditch solutions are usually flawed -- but I think it is inevitable. Addiction to convenience will yield, sooner or later, to the climbing price of that convenience: clogged inboxes, stolen identities, credit card fraud. > > > It's funny how the idea of hypertext, as Nelson foresaw it, was > circulating in the 1960s across the computing community. I recall > discussing it with a systems architect who had worked at Univac in > Blue > Bell, PA, back in the 1970s. We were both doing mainframe work. Back > then, the computing power to do what we were talking about was simply > nonexistent. > > ... > > > > From what I can see, the key innovation that lead to the > Internet we use today was simplicity through layering. I may be > suffering from a layman's oversimplified view of networking protocols, > but it seems to me that internetworking works so well because the > grammar is simple enough to be universal. That's the lesson of > TCP/IP, it seems to me. You're better situated than > I to have a technical opionon on that, though. TCP/IP was a shortcut through ISO/OSI, which layered things beautifully but a bit too abstractly. Long ago there were other layered protocols besides ISO/OSI and its seven tiers, but they all suffered from some degree of proprietary restriction. TCP/IP did at least two things: 1) simplification of ISO/OSI that solved a wide range of problems (as you say), and 2) escape from the proprietary grip of separate corporate interests, into a shared, standards-based framework. The Internet RFC/STD documentation system is a miracle of open-source pooling of ideas -- TCP/IP was immortalized there in a series of RFCs starting with RFC 793 in September 1981. I consider the innovations in documentation, as exemplified in the RFCs, as important as the resulting software itself. But then, I produce documentation for food, so I guess that's my bias speaking... (;-) > > > > > > -- > eric scoles ([email protected] <mailto:[email protected]>) > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "R-SPEC: The Rochester Speculative Literature Association" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/r-spec?hl=en -~----------~----~----~----~------~----~------~--~---
