If Dave can call something the "Yahoo Problem" then I can call it the Dave Winer Problem. It's good way to make sure we see each other's stuff, huh? :-)

Seriously, though... A reporter asked me if I'd comment on his recent posting titled The solution to the Yahoo problem. So I write this without having thought too hard about it or having talked to anyone on the My Yahoo! team about what they think (and some of them go back to the early RSS days and My Netscape).

Dave's proposed solution sounds a bit complicated. In my mind, this has always been solely a client-side problem. The fact is that you click on an orange XML button and the browser does the wrong thing.

In my mind, the obvious solution is to fix the browser(s) or provide a helper app that can do the job if there's no obvious way to extend the browser. I suspect that this would be fairly straightforward for Internet Explorer. Heck, we could do something like it as part of the Yahoo Toolbar. In fact, I've suggested it a few times in the past. The situation with Firefox is even better--someone could build an extension for this. I'm not sure about the less popular popular browsers (Safari, Opera, etc). A helper app would certainly be the fallback position for them.

I'm don't get grok why Dave is suggesting a centralized service that slings OPML files around. Perhaps he'll clarify that need.

Update: Dan reminded me of his Make XML Button Do Something post. He makes the case from a user's point of view.

Posted by jzawodn at January 11, 2005 03:20 PM

Reader Comments
# Robert Sayre said:

The way to fix it is for someone to register a MIME type. application/rss+xml isn't registered, and so RSS is served with misleading MIME types the world over (direct flames about the registration process IANA-wards :). If you do that, you'll find that correct plugin, add-on, browser bar or what have you will get the content. Good things will happen, such as webservers and other software handling it out of the box.

on January 11, 2005 03:27 PM
# Jeremy Zawodny said:

Agreed, a proper MIME type is necessary to make the "help app" idea work.

on January 11, 2005 03:37 PM
# Michael Specht said:

This has always frustrated me about the XML icons, great idea Jeremy. In fact it is such a great idea I am surprised there is not already a hack to do some similar, even without the MIME type.

on January 11, 2005 03:56 PM
# Eric Rice said:

It's funny. Even though I'm not an active My Yahoo! user, I certainly value the function that the AddtoMyYahoo button brings to the mainstreaming of popular weblog content.

How else is every SBC Yahoo DSL customer going to find blogs? Or bloglines? Or feedster? They have to start someplace. And if that Yahoo button is the cornerstone of familiarity, the so be it.

When we were new to blogging, we started with one product and moved to others as we became aware of them. It's a sort of tech evolution.

on January 11, 2005 04:09 PM
# Dare Obasanjo said:

Huh? How is this a client problem? How does a registered MIME type help me in the case of using my mobile phone browser to add a site to my online aggregator like Bloglines, Newsgator Online or Yahoo!'s RSS reader? Currently I can do this today with just HTTP and my dumb web browser.

Michael there are lots of hacks to do what you suggest. I wrote about most of them at http://www.25hoursaday.com/weblog/PermaLink.aspx?guid=385809e7-f9cd-4318-8066-37e0853a97b8

on January 11, 2005 04:10 PM
# Chris said:

Wow, people still pay attention to Dave? I thought everyone just couldn't stand him any longer and stopped paying any attention. Oh well.

on January 11, 2005 04:13 PM
# Anil Dash said:

This is from about a year ago:


No reason it couldn't be for both RSS and Atom.

on January 11, 2005 04:14 PM
# Collin said:

Firefox already has this happening. As you probably know, Firefox displays an icon in the status bar when it picks up an RSS/Atom feed and can add it as a Livemark, but there is a Firefox extension that can hijack that to add it to Bloglines instead of as a Livemark.

It would probably be really easy to modify that code to launch an external protocol, so you wouldn't even need an icon on your page, just the proper link tag.

on January 11, 2005 04:30 PM
# Jonathan Greene said:

Perhaps we'll see something interesting when Tiger ships for OS X. I saw in some Safari screen shots that you can define a default RSS reader... I don't know how it works, just that it becomes a _system_ level pref.

on January 11, 2005 06:45 PM
# Ryan said:

The extension Collin refers to is here:


When installed, clicking on the little orange icon in the status bar will redirect your browser to your aggregator's homepage. I use this to add feeds to my bloglines account. Works like a charm.

on January 11, 2005 07:06 PM
# Mike D. said:

Would you really need a separate MIME-type? What about just using a pseudo-protocol to invoke a helper app? Something like:


That way, whatever application you had on your machine that was designated to handle the "addfeed://" protocol would launch and act accordingly. Firefox could be configured to take over this protocol if the user wanted that to be their reader. Additionally, you could build a hook into Firefox that passed that URL onto server-side aggegators like Bloglines and myYahoo if that was the user's preference.

What am I missing here?

on January 11, 2005 07:29 PM
# Michael Koziarski said:

There's already a feed:// protocol to handle "Subscribe to this url".

If you want to use the firefox auto-discovery, can I pimp my own extension? Feed Your Reader (http://projects.koziarski.net/fyr/).

It allows you to use the 'livemarks' icon to subscribe to feeds in a whole lot of readers. It also allows you to right click and 'subscribe to this link' for sites without link rel= elements.

As far as the 'sharing subscriptions' goes, I'm using the bloglines API.

Dave's post missed the existence of the feed:// uri, but this is understandable as it's not standard.

on January 11, 2005 08:06 PM
# Charles said:

When I read the title "The Dave Winer Problem," I thought you were talking about else. You can't fix that problem.

on January 11, 2005 08:26 PM
# Charles said:

Oops, that was incoherent, due to an editing error. Let me try again.

When I read the title "The Dave Winer Problem," I thought you were talking about something else. You can't fix that problem.

on January 11, 2005 08:28 PM
# said:

Over-reaction. Mention "Yahoo" and "problem" and have a holy jihad launched against you - you are the scum of earth.

Winer isn't the most agreeable of people, but the problem he describes is real, and the solution he offers is just a suggestion. That's one better than ignoring the problem and offering no suggestions.

on January 11, 2005 09:51 PM
# Jesus H. Christos said:

you should take a look at quickSub (http://www.methodize.org/quicksub/whatisthis.html). it 'makes the xml button do something'. (mouse over the orange xml icon on the left side of the page to see it work).


on January 12, 2005 12:40 AM
# Dave Winer said:

I did approach this gently, got no response. I talked with Yahoo people in person, at the Online News Assoc meeting where they were promoting those dumb buttons. Jeremy always makes it personal, I've linked to his blog with respect a dozen times, and do it again today, but he's always biting and scratching. That's the Yahoo technique of dealing with criticism, make sure the person criticizing you is covered with crap, and maybe people won't believe what he says. If they do it to people outside Yahoo, they must do it to each other too. Unfortunately that's not a very good way to develop software.

On the other hand, I use Yahoo for lots of things, I pay them money for services, they run a good server. If they would turn around and try to work with us, seriously, we could do some good.

on January 12, 2005 02:46 AM
# Danny said:

Jeremy, I do think you've over-personalized this - if anything it should be the "Userland Problem" as I believe they were first with "awful arrogant icons".

There are various routes that haven't fully been explored - autodiscovery, mime types, bookmarklets, browser plug-ins and the Control Panel. FireFox's autodiscovery support seems to me the current frontrunner.

Dare's mobile browser could launch a little handler app (preloaded with his aggregator preferences) when encountering RSS mime type which would give him the opportunity to subscribe.

A prerequisite IMHO should be that no-one has to put someone else's icon on their Web page if they don't want to - not even the ugly little orange [XML] one.

I don't think Dave's approach as he described it would offer any advantages over what's already available, aside from the slight potential for vendors to standardise on a (very suboptimal) spec (again).

I reckon putting pseudo-URIs like those of feed:// on the Web is generally a bad idea, I'll call on the TAG webarch document for support.

on January 12, 2005 03:34 AM
# Don Park said:

In my book, the word 'problem' is synonymous with the word 'opportunity'. If Yahoo think this is someone else's opportunity, I am fine with that. After all, success means a lot of work.

on January 12, 2005 03:38 AM
# Morten Frederiksen said:


See also the Syndication Subscription Service: http://purl.org/net/syndication/subscribe/

on January 12, 2005 04:18 AM
# rick said:

This sounds like more buttons to me.

Joe already showed the correct way to do this: http://bitworking.org/news/Atom_Auto_Sub_How_To

on January 12, 2005 06:45 AM
# Jeremy Zawodny said:

Danny: if you don't realize that the title was intended to be funny, please go back and re-read the smiley.

Dave: I asked via personal email and you ignored the question. Who is "us" in this context? Where are the discussions happening? Mailing list? Wiki? Other?

on January 12, 2005 07:49 AM
# Danny said:

Jeremy, right, unfortunately subtleties of verbal communication doen't seem to work in the blogosphere. Whatever.

btw, Yahoo could easily implement the core of what Dave's suggesting by making people's subscription lists available in a machine-readable format. Bloglines does this already. After all Dave only mentions "a server".

But to do it more usefully I think you'd be sharing the subscription data using a P2P protocol of that data between different tools/services (the RDFGrowth algorithm springs to mind). Personally I'd also steer well clear of OPML, it was designed for outliners and is very weak outside of that context.

Sharing data like subscription lists does hold a lot of promise (e.g. Attention.xml, the Semantic Web). But I think the solution to the immediate problem lies along the lines you describe, with client-side handling and (IMHO) autodiscovery.

on January 12, 2005 09:04 AM
# Randy Charles Morin said:

I think almost everybody (sorry Dave and Dare) agrees the solution is the mime-type. But as I found out and mnot before me, registration is not possible [1]. What next?

[1] - http://www.kbcafe.com/rss/?guid=20041120071518

on January 12, 2005 12:02 PM
# Danny said:

PS. I'm putting mine on del.icio.us, whatever anyone says ;-)

on January 12, 2005 12:06 PM
# Phil Wilson said:

Just as an endorsement, but Michael's "Feed Your Reader" extension for Firefox is absolutely marvellous. I've been using it since the first release and it's just really, really great.

on January 12, 2005 03:45 PM
# Dare Obasanjo said:

>I think almost everybody (sorry Dave and Dare) agrees the solution is the mime-type.

You mean almost everyone who doesn't actually write an aggregator thinks the solution is MIME types. Lots of desktop aggregators supports the feed URI scheme. Exactly how many support subscribing via custom MIME types?

One of the major problems with the syndication world is that a lot of the people going about writing specs and proposing solutions aren't actually aggregator vendors. Which is why we end up with stuff like Atom 0.3 and attention.xml

on January 12, 2005 05:06 PM
# name said:

Yes, you must write a wonder of Computer Science like RSS Bandit to understand the issues.

on January 12, 2005 06:26 PM
# Dave Winer said:

Dare, this problem can be solved by the aggregator writers. Ignore the mail list trolls and stay focused on what we can do together to make it all work better. This has to do with users and developers, not flame-baiters.

To Jeremy, I did indeed choose not to answer a bunch of your questions. I know you don't want to work with me, so I don't invest any energy in trying to get you to.

on January 12, 2005 07:06 PM
# l.m.orchard said:

So, I liked a comment Phil Ringnalda had: Why don't we do what MP3 playlists do? An M3U or PLS file contains URLs pointing to streams. You click on a .pls or a .m3u, and streams start in your player. Since they Just Work, I've never bothered to find out what mime types or psuedo-protocols they use.

Why not swap MP3s for feeds? call it a .fss file (feed simple subscription) or even .sfs (simple feed subscription)


on January 13, 2005 11:40 AM
# Andy Todd said:

Problem solved;


Can we all go back to posting cat pictures now?

on January 13, 2005 12:14 PM
# John Furrier said:

I think that Dave makes a point here. In terms of user experience (which Yahoo says they care about) this idea has merit in principle. Yahoo was successful by making a confusing Web simplier. Why not support that same vision now.

I can't recall in recent memory anything Yahoo has done innovative over the past five years. I seems that they never lead any industry initiative. Thats ok from a corporate strategy perspective but don't pretend then throw mud at a guy like Dave who does "lead" and does care about user experience.

Jeremy: I follow your blog with respect and I think that Yahoo has an opportunity here. Dave's idea may be be 'fully baked' but it seems directionally correct to me.

on January 15, 2005 05:57 AM
Disclaimer: The opinions expressed here are mine and mine alone. My current, past, or previous employers are not responsible for what I write here, the comments left by others, or the photos I may share. If you have questions, please contact me. Also, I am not a journalist or reporter. Don't "pitch" me.


Privacy: I do not share or publish the email addresses or IP addresses of anyone posting a comment here without consent. However, I do reserve the right to remove comments that are spammy, off-topic, or otherwise unsuitable based on my comment policy. In a few cases, I may leave spammy comments but remove any URLs they contain.