If anyone had any doubts that RSS is just for geeks, he had better think again.
One sure sign that it's not just for geeks, is when the advertising corporations get their paws on it. InfoWorld puts advertisements into its RSS feeds. No problem there. But if we take a closer look, we see that they also include a web bug, a single pixel GIF that links to a URL with the hostname ad.doubleclick.net and that has an identifier tagged to it.
RSS is going mainstream. And the advertisers will not be left behind.
On January 1 the new CAN SPAM law goes into effect. What can we expect?
An article at CNet news.com provides some insight.
Alan Ralsky, perhaps the most prolific spammer of them all, plans to change his spamming ways, but he will still send marketing messages via email. We expected this. CAN SPAM is all about defining what is legitimate email marketing and what is not.
CAN SPAM is not all bad. At least it's clear what is illegal. Those who decide to break the law do so knowingly. And it is risky to break the law. As Ralsky says:
"Of course I'm worried about it. ... You would have to be stupid to try to violate this law."
I stand by my earlier prediction that eventually ISPs will begin to sell the email addresses of their subscribers. They will do this so that "legitimate" email marketers can avoid the shadiness of email address harvesting. Don't get me wrong: ultimately, ISPs will do this to make revenue. But the email marketers will ask them for the lists, and it will be impossible to say "no."
But I think I can safely make a new prediction. The email marketers will sue any legal entity that gets between them and their end-recipients to filter messages. Remember, the email marketing messages we are talking about are legal, and the intermediaries are interfering with normal commerce. Does anyone remember that Blue Mountain Arts (the company that enabled users to create eGreeting cards online) sued Microsoft over spam filters in Microsoft's Outlook Express?
The news.com article provides further evidence of lawsuits around the corner.
He [Ralsky] would identify himself, as required, and would honor any requests to be removed from his mailing lists, he said. He said that he was counting on Internet providers, in return, to stop trying to block his messages.
Attention ISPs: you've been warned!
To digress a bit, spammers truly are delusional. They are the little guys who are just trying to make it, but the corporations are all against them. My favorite quote from the article:
"E-mail marketing is a good thing," Battles said. "I create jobs. But the media has made e-mail out to be some sort of terrorist plot."
Do you suppose Battles was talking about all the jobs that were created at BrightMail, Message Labs, Postini, and other new anti-spam companies? Mr. Battles, don't you think those people would have been better employed doing something other than cleaning up the spam mess?
Being an Emacs user for many years, I feel Don Box's pain.
Patrick Logan makes some additional points about Emacs vs. IDEs.
I have problems using IDEs. First of all, too many panels get in the way. Too many toolbars, tabs, and status bars are stacked horizontally. I want work space for my editor, and I feel cramped. When I used to write programs on a terminal with 80 columns and 24 rows of text, I used to print source code frequently in order to study it. That all stopped when I was able to view 40 or 50 lines of code simultaneously on the screen. Lots of visual workspace is a major productivity enhancement.
I like using keystrokes for all but the most rarely used tools. The stress caused to the muscles surely must be worse for mouse clicking than for pressing keys on the keyboard. The action is different. Pressing a mouse button does not involve the wrist at all, causing more stress to the muscles of the finger. (Try it!) Typing on a keyboard allows the whole arm and wrist to be used. (Talk to a piano teacher about good keyboard technique and then apply that to typing on a computer keyboard. I myself am a pianist.) So, I believe I am at less risk for repetitive stress injury by using a keyboard for most editing activities.
I also like the fact that Emacs is everywhere I need it: on my primary desktop box, on my laptop, on several other develpment boxes.
For all Emacs advantages, I give up very little. Emacs has many advanced features that I use, many of which are probably unknown to non-Emacs users. Ediff alone is a good reason to have emacs around.
SMTP is the Simple Mail Transport Protocol. Normally, simple is good. But when simple means "too open," we get spam. Spam has driven many individuals, companies, and organizations to start talking about either fixing SMTP, or replacing it altogther. Others are critical of any such initiative, and point to the large installed base of SMTP servers as the reason why SMTP can't be fixed. The critics are right.
However, there are many ways we communicate. Internet mail is just one of them. There are wireline phones, cell phones, instant messaging, wireless text messaging, chat rooms, web forms, weblogs, RSS, VoIP and more. I wouldn't rule out the possibility of even more communications media appearing in the near future.
It's possible that Internet email as we know it may be replaced by other forms of communication, thereby making SMTP irrelevant or possibly even obsolete. This is already happening in a few areas of communication. RSS is being recognized as a way for companies to communicate with their customers and avoid issues with spam filters and spam "noise." Web-based forums are replacing mailing lists. Instant messaging is replacing email for quick messages such as "can you meet me for lunch?"
There doesn't seem to be that much of a difference between instant messaging and email. Instant messages are shorter and more immediate. But it's possible that instant messages could become longer and less immediate. It would be relatively simple for AOL, MSN, or Yahoo to change their instant messaging systems to allow you to send a message to a recipient who is not currently available. Imagine that you log on with your instant messaging client and the client tells you you have messages. A big difference between this kind of asynchronous "instant" messaging and email is that extra control is built in to IM. The IM system verifies the sender and allows you to receive messages only from senders you know. This new form of email -- an adaptation of instant messaging -- could be called "p2p mail."
If spam becomes a problem, p2p mail could become a popular way for individuals to communicate with relatives, friends, and acquaintances.
I don't believe SMTP will ever be "fixed." However, I can imagine that new forms of communication, including some that have yet to be introduced, could make SMTP less popular, perhaps even irrelevant or obsolete. p2p mail is just one example of what could happen.
Microsoft and New York are suing some spammers. Among them is Scott Richter, who was a panelist at the FTC Spam Forum last spring. Richter tried to put on the face of an honest email marketer. Now we see that he allegedly used all the fraudulent tactics that spammers use.
[Microsoft's General Counsel Brad Smith] went on to say that these cases would prove that the defendants in this case used "intentional, deceptive" methods and abused some "514 compromised computers including innocent schools, hospitals and government ministries in the U.S. and abroad."
Scott's bio [PDF] is available from the FTC website.
Scott, they're going to take away everything you own.
Yesterday President Bush signed the CAN SPAM bill. It goes into effect 1 January. There is a lot of commentary around the web, including an article at news.com. Of course, a big part of the law is defining just what is legitimate email marketing. Organizations like the DMA want to know.
I haven't heard anyone discuss what is a next logical step for marketing. Since this is an opt-out law, any company can send you legitimate email marketing messages until you ask them to stop. Since this is legitmate email marketing, the next step is for AOL, MSN, Yahoo, and Internet service providers to sell their list of email addresses to the marketers. There is already a precedent for this practice in telemarketing. The local phone companies sell directory information, and your name and number is included unless you pay them to not include you. Perhaps the ISPs will allow you to pay a few bucks to keep them from selling your email address.
In yesterday's post I linked to an article that profiles a small-time spammer in Louisiana. If we read between the lines, we can see in the article some very, very good news about spam.
Here's a quote:
Fox's days of carefree spamming are past, and so is the good money. She worries that bankruptcy is just around the corner and blames the Internet companies -- who have become more adept at filtering out spam.
So, a small-time spammer is about to go out of business. Sounds like small-time news, doesn't it? But wait. There comes a time in every industry where smaller competitors can no longer make it. The industry goes through a phase of consolidation, where mergers happen, and in the end there are only a relatively small number of large companies remaining. Is this what's happening in the spam business? If it is, then that's good news for two reasons: First, consolidation happens so that costs can be cut through economies of scale. Since spammers already have low costs, how much can consolidation help in cutting costs? My guess is that it can't. So that's good news: spamming may be becoming unprofitable. Second, if the spam industry consolidates down into a relatively small number of spam companies, that means fewer targets for law enforcement.
This is good news indeed.
From time to time articles are published on the web that give us a glimpse into the operation of a spammer. In every case, the spammer expresses that he feels justified -- that he is a legitimate businessman -- that he provides a valuable service to customers -- that he is no different from telemarketers or other marketing firms.
The reality is, that spammers create more economic damage than economic value. They can be compared to a company that employs people and sells legitimate products -- good economic value -- but also trashes the environment -- bad economic value. If a spammer makes $10,000 per month, but causes $100,000 per month in damages to others, that's a business that can't be justified, no matter how you spin it.
Maybe one day one of these reporters, when writing a story on the operations of a spammer, might just ask the spammer how he feels about the fact that he creates more economic damage that he does economic value.
Footnote: See Spam wars play out across Internet from the Atlanta Journal-Constitution, which profiles a small-time spammer in Louisiana.
Greg Reinacker discusses the issue of how RSS should be used to read posted content. (Greg is the author of the NewsGator RSS aggregator.) There are two issues: The first is should RSS aggregators also grab the referenced content. The second is what should be put into the RSS entry.
On the first issue, it's clear that if users want an aggregator to grab the complete content, someone will create a news aggregator that does just that.
There are practical problems, though. Grabbing the content involves HTML scraping, which is imperfect at best. There are countermeasures against that kind of activity, too. An author could divide the story among several pages linked together, so that HTML scraping would grab only the first page. HTML scraping could also lead to a lot of worthless HTML, images, and style sheets being stuffed into a news aggregator.
On the second issue, news aggregators are great for offline browsing. When I first got a Handspring Visor (PDA), I really liked AvantGo. I used it to read some of my favorite news sites while at the mall or elsewhere with the family. RSS is an AvantGo-like service. It's perfect for PDAs or tablet PCs. AvantGo is nothing more than a news aggregator. RSS does what AvantGo does.
There are reasons why a content author might want the RSS feed to be a teaser. The primary reason is economic: if a web site is supported by advertising, then certainly the content author wants you to view the web page decorated with advertisements, and not a stripped down version in an RSS feed. Advertising-supported authors could still put complete content into RSS feeds -- they could include ads in the feed.
There are also reasons why a content author might want the entire content in an RSS feed. AvantGo makes money by allowing corporations to use the service to disseminate their own internal information. People within the corporation need to synchronize with information that is timely. RSS could serve those users. RSS could also serve the needs of bloggers who post ideas and want maximum exposure for those ideas.
On the reverse side, the readers may or may not want complete content included in the feed. For offline reading, having the complete content available is necessary. For online reading, having only summaries presented is a great way to scan a vast amount of material efficiently.
What is necessary is a way for both summaries and complete content to be made available. Complete content could be optional both from the author's side and from the reader's side: complete content may not be provided in the feed, and readers may elect not to grab the complete content even if it's there. Perhaps the necessary changes will be adopted in Atom. (I admit to not following developments in Atom.)
With complete content being optionally made available in feeds, I think great things could happen. First, many users could take advantage of feeds as a means for synchronizing content. For example, important company information could be made available via private feeds in order for mobile users to synchronize and have up-to-date information on the go. Second, this opens the doors to more efficiency in the dissemination of information -- even ad supported content, by including ads in the content. Offline reading is just one possibility. There is also the possibility of dissemination through peer-to-peer networks. And finally, before us is the chance to focus on information, rather than entertainment. It seemed that during the dot-com boom years many viewed the Internet -- and the WWW specifically -- as a medium not much different from television. Information workers, though, have always had a different view of the Internet as an excellent tool to help them accomplish their jobs more efficiently. The move to XML and well-formed content, beginning with RSS/RDF/Atom, may not have entertainment value, but it certainly benefits information workers.
Jeremy Zawodny has posted predictions for 2004. I think the predictions about personalized search are interesting. I'm not sure what Jeremy is thinking about personalized search, but I have a few ideas of my own.
What if I published a list of sites that I trusted, and search listings where ordered by "distance" from the list of sites I trust. I'm thinking about this more in terms of relevance than in terms of trust. The sites on my trusted list are relevant to me. If those sites link to other sites that are close in terms of subject matter, then the sites that are "closer" to me should be more relevant. At least that's one theory. Such a scheme also has the advantage of being "decentralized," in the sense that everyone will have a different collection of trusted sites. That would make search engine spamming more difficult.
Another big improvement in search engine technology would use feedback from Listing Page N to improve the relevance of entries on Listing Page N+1. Instead of links at the bottom of Page 1 to take you to Page 2, Page 3, and so on, there would be a link next to each entry in Page 1. If I didn't see a relevant entry on Page 1, I would click on the 'next page' link next to the entry that was the most relevant. Hopefully, that feedback to the search engine would assist it to display more relevant entries on Page 2. Feedback from Page 2 would yield still more relevant entries on Page 3, and so forth.
I'm getting back into SIP -- the session initiation protocol -- and at the same time I'm learning about JXTA -- the peer-to-peer framework. I'm seeing a lot of similarities. JXTA provides abstractions: peers, peer groups, rendevous, routing peers, and so on. So, as I think about the abstractions, I also think about SIP, and I can see how SIP fits the abstractions identified in JXTA.
SIP is a peer-to-peer protocol. So, it's not really a coincidence, but it certainly is a positive testimony to the clear thinking of JXTA's designers. A SIP User Agent (UA) is a peer. SIP messages are advertisements. A SIP proxy server is a router peer. A registrar is a rendevous peer.
SIP is complex. JXTA is relatively simple. I'm not exactly sure why that is so. SIP is used for call control in IP telephony, and call control in the PSTN is complex. But I have to wonder if SIP would not have been simpler if it had started with a model of peer-to-peer communication like that in JXTA. If I were to start work on a new peer-to-peer protocol, I would consider building upon JXTA.
It should be interesting to learn more about SIP and JXTA and to see the extent to which they are similar or dissimilar.
Dividing computer programs into "procedural" or "object-oriented" is a false dichotomy. Procedures are how we get things done. Without procedures, how can we do anything? Even object-oriented programs must use procedures.
As an examle of this false dichotomy, see Understanding Object Oriented Programming.
So, now I ask: How can I write a recipe -- say, for baking bread -- as an object-oriented recipe?
My point is, that we tell beginning programmers that writing a program is writing instructions for the computer to execute. Many programming primers compare a computer program to a recipe. If procedural programming really is fundamentally different from object-oriented progamming, and if object-oriented programming is preferred, shouldn't we have a different way to teach beginning programmers? Can we show them an object-oriented recipe?
There's an article at Wired News that provides a few more details about Yahoo's plans for "fighting spam."
Tim Bray writes about coding in general as an over-X-years of age programmer and about C coding in particular. I feel I could have written the first several paragraphs myself! There does come a time, after you've written code for doing ${task} for the umpteenth time, that you feel it's time to move on from writing code. It's true, though, that experience does matter. I feel my own code is much better than it was ten years ago.
Rawdog is an RSS Aggregator Without Delusions Of Grandeur. I'm trying it.
Written in Python, rawdog keeps track of your subscribed RSS feeds and creates a webpage with the updated contents of the feeds.
I installed rawdog on my hosting account and set up a cron job to run rawdog every 15 minutes. You can view the generated page. I'll be adding more feeds daily for a while. What I need now is a simple CGI script to add and remove feeds.
As reported by InfoWorld, Yahoo thinks it has found the Ultimate Final Perfect Solution To The Spam Problem (UFPSTTSP). Wait -- don't yawn just yet... Well. Okay, go ahead and yawn.
I can't find any details on their plan. As explained in the article, this is a trusted sender scheme based on public key cryptography. It uses the domain name system, but no details are given. Does it create new DNS records that contain the public key needed to verify that domain? Does everyone who participates have to get a certificate signed by a CA?
A system that stores the public key in a DNS record is similar to the simple sender authentication that I had been thinking about earlier.
What exactly does a trusted sender system do? Spammers register hundreds of domain names. So exactly how does a trusted sender scheme stop spammers? It's a rhetorical question. The answer is: it doesn't.
But, let's look at the bright side of this. If it is a simple scheme of adding a public key to the DNS to be associated with the sending domain, then that seems perfectly acceptable to the little guy, like the guy who wants to run his own mail server at his home. You would need to add the public key to the DNS records and you would need to use the private key to sign messages you send. Then, maybe AOL wouldn't have to take the draconian measure of blocking mail from any consumer Internet account.
Perhaps a scheme that permits the mail submission agent (MSA) to sign the message headers isn't such a bad thing. The question is, will such a scheme ever achieve critical mass? If it does, that would be interesting, in the sense that it took a company initiative -- and not the IETF -- to bring it about. The IETF has shown itself to be paralyzed in any effort to solve the spam problem.
Mark Shuttleworth has an interesting story. He founded Thawte, the company that sells relatively low-cost web site certificates, and became exceedingly wealthy when Verisign acquired Thawte. Mark paid $20M, if I remember correctly, to fly in space on a Russian spacecraft. Now he is funding new projects, both commercial and open source.