Microformats are perfect for SPAMMERS
I’m very worried about the use of Microformats for contact details and I certainly won’t be using them.
According to Microformats.org, the hCard is a simple, open, distributed format for representing people, companies, organizations, and places.
The purpose of the hCard is to make it easier to download a person or company’s contact details into an application such as Outlook. They’re machine-readable and machine-discoverable.
Why Microformats are good
At a glance, they’re brilliant and the hCard is good because it means no more copying and pasting text from Web pages. Furthermore, Microformats is a decent example of how the Semantic Web can benefit everyone, albeit a very tiny example and a tiny implementation of the Semantic Web.
Why Microformats can be bad
So, if its now very easy for tools (machines) to find your contact details online, won’t this make it even easier for SPAMMERS to find your details and automatically load them into their contact list?
I tend to use the format paul [at] segala [dot] com to make it a little more difficult for SPAMMERS such as easyJet to email me and I still get hammered with loads of rubbish. So, I’m certainly not going to put my email address in a machine-readable format to make it extremely simple for them to grab my details. Furthermore, I’m not going to use a Microformat for my postal address so direct marketing companies can have a field day.
Think about it, if everyone had a hCard for their contact details, SPAMMERS and direct marketing companies would no longer need to employ humans.
Perhaps if we coupled the hCard with Friend of a Friend (FOAF) we could enable trust in some way – i.e. so only trusted parties could download your hCard.
Marc Canter thinks Microformats should be even more open – I agree as long as it’s not the hCard.
Have I got this totally wrong (which is likely), or is everyone else jumping on the ‘I love Microformats and OpenID’ bandwagon in case they’re not perceived as ‘with it’?!
13 Responses to “Microformats are perfect for SPAMMERS”
Leave a Reply
[...] Paul Walsh (of Segala) asks: “Have I got this totally wrong (which is likely), or is everyone else jumping on the ‘I love Microformats and OpenID’ bandwagon in case they’re not perceived as ‘with it’?!” – Two different issues here dude. 1) OpenID allows for federated IDs, for SSO (single sign-on) and hopefully (soon) attribute exchange. That’s a good thing. And easily spam resistent. 2) Microformats on the other hand are a partial solution to ‘what are the meta-data standards’. That of course is a good thing to solve, but without a complete solution – it’s – well only partially solved. So I’m not asking for more OPEN microformats – I’m asking for microformats as feeds and file formats not JUST page tags. Get it? [...]
I am not an expert on this, but I recall having the same issues with the FOAF format – machine readable contact details.
Thanks for dropping by Mike. You’re right, you could have the same issue with FOAF at present. The FOAF specification is still under development (sigh!). So, some trust elements could be added. The w3c Content Labels working group is lobbying to have it updated to describe organisations as well as people.
[...] I just noticed that Marc Canter has responded to my original post about Microformats and OpenID. Marc reminds me that they’re two different issues. I actually knew this already so I should have known better and not mentioned OpenID in my post. Although, it did attract Mike’s attention so that’s a good thing. [...]
It’s important to note that FOAF or hCard by definition do not need to expose all data elements (property values). When a service provider generates FOAF or hCard they must provide their customers/members with the ability to designate what stays private and what goes public. The Semantic Web presents immense opportunities to Social Networking Service providers once they are able to understand the implications of my comments above since they are all contact management service providers at the end of the day with SPAM protection as essential part of the value they offer.
I note that already does this re. FOAF.
Thus, a Social Network can provide Open Data Access (via FOAF and hCard for instance) and continue to grow. The real kicker is that growth will actually be exponential since the value consumption will feed of machine to machine consumption of basic and value added services instead of the human to machine dominated interaction of today.
As I said over at Tom’s blog, if you don’t want your information to be used by the wrong people, don’t put it on the web. hCard makes everyone’s lives easier, both the good guys and the bad guys.
Spammers overcame silly messing like email obfuscation years ago. I’m happy for all my details to be out there as a hCard and I’ve seen no increase in email/postal/phone spam since I did.
Conor,
Again, there is no rule that says you put data in hCard or FOAF that you don’t want to be exposed to the public. The key thing is that the hCard or FOAF generator honors this.
For instance, you can build a valuable data source from a social network without revealing email and other contact details. The richness of the graph the underpins the social network will naturally stimulate the need for more personal connections to be established etc. And it is at this point that the hCard or FOAF provide can kick in as the communications processing manager based on in-built message handling and routing rules.
Social Networks and contact manaement engines at the end of the day, whether they know this or not.
Social Networks that play well with the Semantic Web (along the lines I describe) will provide valuable White List Services to a new generation of Rules Bases MTA (Mail Transfer Agent) filters and Sink Drivers (the piece that persist mails on behalf of the MTA into a file system of DBMS engine).
SPAM will be killed by a combination of RDF Data based White Lists, Rules, and MTA filters and Sink Drivers.
Yeah, I made the point about “how much” on Tom’s blog. I think the only field that is compulsory in a hCard is fn (full name).
Paul, we all agree the semantic web is the way forward. One way is microformats and the other way is RDF.
The link below is from my friend Mark Birbeck who co-wrote part of the XHTML2, Xforms and RDFa specs showing there is a possible third way RDFa.
http://internet-apps.blogspot.com/2006/10/rdfa-and-microformats.html
Microformats are working today because they are simple to implement and are providing immediate value already. Tails, Operator and WebCards along with sites Flickr, Yahoo, Upcoming etc
RDF is the logical systematic way to create a search and query internet but it feels like it is taking for ever to get started and is so hard to grasp for the average user with no immediate return.
The comparison I will offer is the fight of HTML against the Client/Server C++ world.
I remember when at Netscape presenting the idea of a browser with simple HTML and being told by corporates it will never take off it was too simple etc. Well we all know what happened there.
I think we are going to see a convergence of semantic ideas in the next few years thanks to XHTML2 (namespaces), GRRDL and browser support for Xforms etc.
So my interest in MF’s now is because I can do something with them and they are adding value. I am looking at RDFa and XHTML2 and waiting to see how much support Firefox3 adds.
The one company that I think will come from left-field to make RDF work is Google. They are big supporters of Atom/GData/XQuery, Atom Publishing Protocol which is in the new Blogger release to replace XML-RPC. Add to this their support for XMPP in GTalk and GCalendar and you can see that Google could start to add in FOAF support firstly in Atom. Google have not supported MF’s which makes me think they will leap frog everyone and go straight to the RDF semantic web and make the web both searchable and easy to query. see Gdata.
So it is actually a semantic point whether we trust hCards or FOAF.
Sam,
Microformats vs. RDF is not a battle that anyone wants to wage. These efforts are so complimentary I cannot put into words.
Microformats provide a nice foundation for unobtrusively producing RDF Data Model Instance Data without placing the burden on the end-user. In short, this is another effort (like Web 2.0 as a whole) that actually helps to crystallize the entire Semantic Web vision.
Microformats and RDF are good for each other. Mark Bierbeck certainly knows that hence his effort re. RDF/A
.
I will soon be exposing some live demos to expose the critical point I am trying to make.
I don’t think it’s a competition between Microformats and RDF. In fact, I don’t think you can really compare them. When I chatted with David Sifry from Technorati in Paris his face almost turned red when I mentioned RDF and FOAF. Personally I think it was the usual misconception about the Semantic Web that prompted his retort. Although he did give me Tantek’s email address when I gave examples such as Content Labels as a means of demonstrating benefits in the very near future.
I don’t believe the Semantic Web has to be all about RDF either. Further to your question about Firefox (Sam), it’s a huge supporter of RDF already, so introducing new concepts such as Content Labels is seamless.
[...] I just noticed that Marc Canter has responded to my original post about Microformats and OpenID. Marc reminds me that they’re two different issues. I actually knew this already so I should have known better and not mentioned OpenID in my post. Although, it did attract Mike’s attention so that’s a good thing. [...]
I’d like to confirm what other commenters have mentioned: hCard has only two required properties, vcard (the container) and fn (formatted name). Authors therefore have granular control over what data they want to be machine readable. I regularly use hCard to mark up my name and URL only.
I also know there is a blogger (can’t remember who, unfortunately) that uses server-side technology to control who can view his contact info. Everyone gets his name and URL, but only friends who log in with their OpenID get his full details.