Google and BlogThis
A number of people, some of whom should know better, have been commenting about Google adding a BlogThis button to the latest beta of their toolbar, claiming that it’s somehow shocking and unGooglish. Not so.
For one thing, it’s completely Googlish: search for something at AltaVista, and they offer the option to repeat your search at Lycos. Lycos then offers the option to search at HotBot, AltaVista, Overture, or their copy of DMOZ. Go to HotBot, and they offer branded results from FAST, Google, and Teoma. Search at Google, and if you want results from another search engine, you better search for “search engine”, because they aren’t going to lead you by the hand to their competitors.
But even if Google did decide to encourage people to use competitors to their Blogger service, it’s not that simple. Note that although I have absolutely no incentive to encourage people to use any particular blogging tool (I’ve probably got more code investment in bBlog than any other tool, and I think that’s my first link to it), thus far I’ve done a Mozilla/Firebird extension to add right-click BlogThis for Blogger, and nothing else, despite the fact that I myself mostly use Movable Type, and only rarely use Blogger anymore.
The cases are exactly parallel: for both Google and for me, adding Blogger BlogThis requires the three lines of Javascript to get the selected text, current page title, and current page URL that are common to all BlogThis-like scripts, and then window.open("http://new.blogger.com/blog_this.pyra?t="+t+"&u="+u+"&n="+n). Full stop. That’s what it takes, for every Blogger user.
Doing the same thing for locally-installed blogging software involves asking for the path to the script that accepts BlogThis-like URLs (or the weblog URL, and hoping to find an RSD file linked there that offers something useful), and then for MT asking for preferences about which features (extended entry, allow comments, etc.) to show in the popup, and for b2 and friends asking which fork (assuming you know about the URL pattern for every fork). For Radio, it involves making sure the user has Radio Express installed, and then finding out the path to it.
The alternative is to use one or several blogging APIs. That involves a whack-load more code than just five lines of Javascript, and to offer a fairly good experience involves supporting the Blogger API for Blogger, metaWeblog for Radio and probably several others where that’s the best API available, and metaWeblog-plus for things like MT that expose more through their own API extensions. The end result will be fragile, since people have the unfortunate habit of changing APIs, and in almost every case it will be a worse user experience than using a BlogThis-like “API”, because when you BlogThis you get the application’s own interface, compressed down into a popup, while with any API you get someone’s approximation of some of what the interface offered at some point in the past. A standalone API-based program sometimes makes sense, for writing posts offline, for avoiding browser crashes, or for working around the missing features in some blogging programs, but for launching from a browser toolbar none of those fit.
So Google has three choices: support their own product through the best possible interface, support a fixed list of products through a bunch of application-specific code and confusing preference UI, or do a fairly poor job of supporting anything and everything that works with any published API. I don’t see anything the least bit surprising about them having picked the first choice.
Update for scripting.com readers: In case this isn’t quite clear, I’m quite capable of finding out whether or not a given program supports the Blogger API, and in fact I know without looking that Radio and Manila both do. My point is that a “Blog It With Anything” button based on the Blogger API would be a worthless curiosity, because it wouldn’t be able to implement enough functionality to meet the needs of, well, pretty much anyone, since even Free Blogger now allows for titles, and the Blogger API doesn’t. Throwing in a BlogThis button probably took three minutes, and provides the perfect interface for Blogger users; doing a Blogger API program and testing it would take between hours and days, depending on whether it was someone who already knew their way around XML-RPC and the Blogger API, and would provide a truly terrible interface for pretty much everyone except people who don’t use post titles, and don’t care about fancy stuff like Textile or TrackBack or Radio’s Link field. Like the bulk of the things that were developed for the Blogger API (and for that matter, most of the things developed for metaWeblog, too), it would be cool once, until you noticed that you had a bookmarklet a quarter inch above it that would do the same thing, vastly better.
FYI, Radio supports the Blogger API, as does Manila. It would make perfect sense, imho, to limit the Blog This functionality to the Blogger API, and to have the values default to the Blogger host.
There is a difference between ’default values’ and ’hard coded values’
Just becasue Radio, Manila, bBlog, whatever, supports the Blogger API, doesn’t anything, becasue the GoogleBar only supports Blogger.
All of my MT blogs, all but one of my Blogger Pro blogs, and my Radio test blog, all use titles. The Blogger API doesn’t, so as far as I’m concerned it’s useless. I’ve written to it exactly once, in some private code that was supposed to get draft answers for my Blogger FAQ, where I would copy-and-paste the titles to the actual title field before publishing. For anything else, a prominent ”feature” that would let me post without a title would just be a prominent nagging annoyance.
In a BlogThis! popup, I can add a title (and a separate URL for <link> if I’m so inclined), set the post as a draft, change the date and time, spellcheck, or upload a file to go with the post. In an MT It! popup, I can choose whether to send a TrackBack ping, add a title, an excerpt, keywords, and an extended entry, choose a category, decided whether to allow comments and TrackBack pings on the entry, and decide which text formatting plugin to use.
My point isn’t who supports what API, it’s that unless for some reason you don’t want to have a browser open (or you’re still using IE, so you expect it to crash frequently), if you use any API other than one designed specifically for your program, it’s going to be missing some of the things you want to use: for me, if it doesn’t have titles, excerpts, and send TrackBack pings, it’s useless: I need all three, and I’ve got a bookmarklet right there above the Google toolbar that will. For a Blogger Pro user, it’s likely to be the titles, and maybe the option to set a post as a draft. For a Radio user, it’s likely to be the absence of the wizzy editor.
In every case, the native interface is vastly better than some lowest common denominator ”just a textarea and a Post button” API version. I’d much rather see people work on a way of advertising the URL to their BlogThis-like implementation, with a standard set of query string varibles and a way in the program to set defaults for extra features when the URL doesn’t specify which ones to use, so that any program could just ask for the weblog URL, use that to autodiscover the BlogThis URL, and know that adding u=url&p=posttitle&t=text to that would bring up the user’s favorite version of their native interface, than have any API that lets me do complete posts, whether it’s in XML-RPC, SOAP, or POSTing the raw XML. That stuff’s all interesting to spec designers, but I just have one desire: get the URL of the page that’s up in my browser right now, along with the text I selected, into my blogging program, where I can turn it into a post.
Also, is this post in response to my article? If so, why no link? If not, then why not a link to the people who ”should know better.” When did your blog become a moral play? I liked it better when you stuck to the facts and pointed to the things you’re commenting on.
Compare:
””When did your blog become a moral play?””
Contrast:
””When did you stop beating your wife?””
Actually, I started out beating on her, and never stopped: I’ve always done that sort of post, and probably always will. Even though some of my readers are people who are just starting out with blogs, my ”target market” is people who already read the stuff I’m talking about, so I quite often do comment on this sort of distributed, unthreaded, generalized ”conversation” without linking into it. Especially when I’m objecting to it. In this case, what tipped me over the line to posting about it was someone (else, not Dave) that I respect going waaaay over into tinfoil hat country, making it extremely personal, in a post that I would never in a million years link to, and so I decided not to dig around in other people’s archives for links to posts that I know everyone has already read.
And since for some reason I lost a point of PageRank in this last update (poor, pitiful me), a link from me is even less valuable than it was, so nobody’s really missing out.
Calm down Dave. The world doesn’t revolve around you.
As they say in Yiddish — Oy Gevilt!
Such angst. Yes, Mr or Ms Anonymous got it right. I wasn’t really asking when exactly his blog became a morality play. How dishonest of me (sorry for the sarcasm).
And to Mr Eisenman, thanks for the reminder.
Y’all take care and have a great day.
Despite http://www.voidstar.com/node.php?id=1405 I can’t really get worked up about this.
Add a bookmarklet to a links toolbar or add an IE hack to the right click menu or add a button to the Google toolbar, what’s the difference? If Google had used the blogger API then I’d agree with Dave and I’d expect them to support multiple endpoints. But all they’ve done is add a bookmarklet button to their toolbar. Every major blog tool has a bookmarklet available but they’re all different. It’s very hard to see how Google could have allowed the user to change to another one.
I understand the criticism but I can’t see how they could have done it differently. At least without taking it to the next level.
Julian do you see where this is going?
Lots of confusion and gated communities. This is the beginning.
It’s not much of a technical question, either way would be easy, it’s a matter of millions of toolbars. Suppose another blogging tool did a deal with Yahoo, just hypothetically. Should they exclude Blogger? I wouldn’t want to do it, I would want to use that moment to grow the whole concept of blogging, and set the bar really high for the competition, because it would be easy for users to configure the toolbar to talk to some other weblog software.
Make the products compete on features and performance, not ignorance of the users. These lessons were hard-won in the browser wars. You can see the same issues coming up again. Now is the inflection point, the time when the little people, you and me, can set the tone for the times to come. Later we’ll just be able to piss and moan about what could have been.
My opinion, of course, ymmv, ianal, my mother loves me, etc.
I note that saying something isn’t surprising is not the same thing as saying it’s good.
I still don’t get it. It is Okey for Google to support only Windows, but not to support only Blogger (which it owns)?
To support only Internet Explorer on Windows: yes, I’d much rather have people getting excited about the way that the only access to PageRank on a toolbar is to run IE/Win. Here, there’s a million other ways to do it, so having the BlogThis button is just a bit of advertising; there, the only way to access a piece of information is to run a particular browser on a particular operating system. That is evil.
Supporting only MS/IE is of course evil. ;-) It’s also real-politique, just ask Mac/Linux users. I would love to have somone hack the Google and Alexa proprietary toolbar bits and produce a Mozilla Firebird extension that duplicates them, but I don’t expect it any time soon.
But mostly I’m puzzled by what the detractors would have prefered Google to do.
– Build a full blown blogger API client into the toolbar? Hmmm. Now there’s an idea.
– Offer a user customizable button where you could put *your* bookmarklet? All the major browsers already have this. Why duplicate it?
Yeah, it’s too bad that the wonderful efforts of the Mozilla Googlebar folks have resulted in a toolbar that virtually duplicates the IE/Win version, except for PageRank, which is the only reason I ever use it. And I suspect without having asked around that any attempt to hack PageRank requests would bring the full force of Google, Inc. down on your cringing head. So I don’t think I’ll try, even if I knew how to get started (sniffing a whole lot of packets coming from IE/Win, I suppose).
A brief session with ethereal, and the google and alexa toolbars suggest that both do simple http GETs. Alexa returns an XML formatted block of data while Google returns 19 bytes of data containing the page rank in plain text.
This shouldn’t be too hard. I wonder if the DMCA has relevance here?
You might find these interesting
http://tinyurl.com/gq4t
http://tinyurl.com/gq4u
http://tinyurl.com/gq5a
I have an idea, Dave. Why don’t you create your own toolbar that’s as useful as the Google Toolbar?
Mark, I could, but that’s hardly the point.
My toolbar of course would allow people to post to any blogging tool, including Blogger. And if it’s very good a few thousand people might use it. I doubt if it would make much difference to Google, or to the world.
On the other hand Google told us they wouldn’t treat Blogger differently. I doubt if they foresaw this at the time, but it seems to apply.
Dave, your arguments would be a whole lot more convincing if the syndication tools you’ve worked on supported RSS 1.0, and not just your own version. Or prove my cynicism wrong – build us the toolbar!
Danny okay, got it. I don’t think your feature requests are really on-topic here. Also, as far as I know the UserLand aggregator does indeed work with RSS 1.0 feeds. Nice try. ;->
”Blog This” is just a smokescreen. Seriously, what percentage of Google Toolbar users also blog? Most people don’t have the time to feed their egos by posting random crap on blogs (like this comment, for example).
If Google really wanted to be evil, they’d have a ”Subscribe to This” button that uses auto-discovery to subscribe the user’s aggregator (perhaps owned by Google?) to whatever blog the user is viewing.
The RSS crowd rants about ease of publishing via ”Blog This,” but the real killer app is ease of subscribing. The average user is confused by all those myriad XML icons. (Dont’ believe me? Recommend an RSS feed to a non-technical discussion list. 80% of the responses will be, ”I clicked on that XML thingy and saw all this weird stuff.” The confusion about how to subscribe to a feed is very widespread.)
If you want to encourage open standards for the masses, focus on the consumption side of the equation, not production. I propose something like this. And if you do this, publishers will love you forever. Accurate subscriber counts. Hallelujah!
I sent this to Sam Ruby via email last week, but he never replied. I sent something similar to Dave a while back. No response there either. I guess they’re both too busy arguing over the crucially important (to Google users?) namespaces vs. core elements issue.
Dave, I wonder how you would react if someone came along and told you how to write your blog. In fact, I guess I know because you used to allow comments on Scripting but turned them off when the heat got too high. This place is Phil’s to do with as he sees fit and I don’t see how you, of all people, should be calling him on the carpet for it.
Bill, people tell me how to write my blog every day, many times. It comes with the territory.
Derek, I like the idea you sent to Sam, wish you had sent it to me (although there’s not much I can do about it right now).
About the idea you did send to me, I think I responded, we already have that, it’s the Web Bug Simulator, unfortunately the only aggregator that supports it is Radio, but it gives you a pretty good idea of who’s subscribing to what.
If I didn’t respond to you I responded to someone who sent me the link to your piece. Sorry I get so much spam that unsolicited email from people I don’t know usually gets lost. But I did get a pointer to your piece, and did think it was a good idea, I just done it over a year ago.
This raises another point that came up in a lengthy phone conversation with Jon Udell yesterday. We have so many solved problems that have been written up and implemented in one or more tools, but people don’t trip on them, or not enough people know about them to get them on the radar.
I fully agree that we spend too much time revisiting problems that were solved many years ago, digging them up and arguing over breakage, as if that were an option. Honestly it isn’t. The only thing that can change is there can be more confusion. I try not to invent new things. Danny Ayers of course is already busily researching proof that I’m a hypocrite. Oy.
I think we need one of those Cones of Silence they used to have in Get Smart. You have to be pretty old to remember that. Unfortunately I am that old.
Have a nice day.
Cone of Silence – I remember Get Smart, but only the reruns.
Re: WBS – I came accross that in my research a while back, and it’s a step in the right direction, but it only works to track whether an item has been viewed (i.e. loaded in a browser), not whether someone has subscribed to a feed and deleted the item without reading. That may seem like a small distinction, but it’s actually quite huge in publishing and marketing, where ratios are at least as important as absolute metrics. (Ask anyone who’s ever fretted over their abandoned shopping cart rate.)
And I agree that the RSS debate cum flame war is spending waaaay too much time solving potential problems (wherein an engineer asks ”what if I someday want to sort/search/pingback/whatever based on attribute <foo>”) than real ones. I suspect if you could show a detailed subscribers vs. open report to the circulation people at BBC and the New York Times — the circulation people, not the engineering team — the budget purse strings for for RSS-related activities would loosen up quite a bit. And I don’t think they would ask a lot of questions about standards. They would ask about business metrics: How many people are subscribing? What percentage are clicking through? What happens after they click through?
Answer these questions and the people who pay you money (which are not the same people who download Google toolbars) will not give a hoot about standards wars.
It may sound like crass capitalism, but investment by companies in a technology does far more to promote its becoming a standard than anything Google or Microsoft can do. Remember Blackbird? It died when Microsoft realized that companies were already investing in HTML. Get companies to invest in RSS and the rest will take care of itself.
Derek,
The BBC currently generates all the links in their RSS feeds to pass through a click-through counter which contains both the article to read and the version of RSS that the click came from so they’d have very accurate stats as to what is going on.
This seems like a very sensible way to work to me – no client smarts needed, the stats are complete and would be counted in the same place as their normal web stats I imagine.
Derek,
Here’s a bookmarklet that will auto-subscribe you to the feed of the page you’re looking at, if it supports the RSS autodiscovery LINK tag:
http://diveintomark.org/projects/misc/autorss/radio.html
This only works for Radio Userland, but it can be adapted for other browser-based aggregators like AmphetaDesk as well.
Or non-browser-based aggregators: SharpReader and lots of others listen on port 5335, mostly looking for clicks on ”Subscribe in Radio” icons, which brings up a problem we’re likely to have when aggregators start registering as content-type handlers, rather like the existing problem I have with port 5335 listeners. I usually have around a dozen aggregators installed, and for the most part they are truly awful about coexisting: rather than trying to grab 5335 and failing (quietly or noisily, depending on how important it is to them), them mostly crash violently, or hang, or (my least favorite) sit invisible, in limbo, waiting to finish loading when and only when the port they want is available. Based on that, I’m guessing when they start registering as content handlers, we’ll see a lovely repeat of the way that things like Winamp, Real, and WiMP viciously compete to handle audio and video files, stealing the registration, offering to steal it back whenever it’s stolen…
Phil: Based on that, I’m guessing when they start registering as content handlers, we’ll see a lovely repeat of the way that things like Winamp, Real, and WiMP viciously compete to handle audio and video files
I’ll take that any day over the status quo – a bunch of confusing icons that lead you to a bunch of gobbledygook (to the average user).
You should check out the Syndication Subscription Service.
It attempts to solve both of these problems – using a generic subscription button/link, and a local listener that responds to requests like the ones from ”subscribe in radio” buttons.
Disclaimer: I made this.
Peter – If that is true, then bravo. But if I understand you, it doesn’t count subscribers, just clicks and opens. Talk to an email newsletter publisher. The open rate is as important as the number of opens. The click-through rate is as important as the number of clicks.
Paul, a bookmarklet is nice, but it only works with browser-based aggregators. Should the average user – e.g. my mom – expect her Hotmail email to work differently than her Eudora email?
Like I said, users see these XML icons and don’t know what to do. Which is why I originally suggested that aggregators register as the default handler for application/rss+xml and publishers start using this as their content type. Surely if publishers can bother to implement auto-discovery, they can go the extra mile (i.e. one line of code) to issue a new header.
It’s amazing to me that this still hasn’t happened. Real Networks, Quicktime and Windows Media all respond the same way when I click on, say, a file of type application/mpeg. Can’t aggregators get their act together enough to do something similar? Why all this right-clicking, bookmarklet and href-to-localhost nonsense? IMO this is where the most important work needs to be done vis a vis standards, Bloggers will figure out how to deal with ”Blog This,” it’s the consumers who need something simple and predictable.
And if we don’t fix this, well, then don’t be surprised when Google/AOL/Microsoft ”Subscribe to This” doesn’t behave the way you’d like.
PS I just got back from speaking at Rocky Mountain Internet Users Group (along with Chris Locke) about RSS and blogging. Explaining how to subscribe was the trickiest part. I showed them what happens now when you click on an XML icon and everyone laughed.
Derek,
I think getting at that info should be doable the way the BBC are doing it. The combination of hits on the base RSS file and the click through from distinct IP’s would give good numbers – given proxies adding the correct headers that have the requesting client IP listed.
The method they are using has the advantage of showing up in their normal web stats.
I was impressed when I noticed it not so long ago – was looking for a way to get those kinds of stats myself.
Peter, what about multiple accesses to a feed by the aggregator (i.e. once an hour throughout the day)? In my tests, many aggregators don’t honor eTag/if-modified-since. And even when they do, my guess is that most web stats packages would still call it a page view.
>>but it only works to track whether an item has been viewed (i.e. loaded in a browser), not whether someone has subscribed to a feed and deleted the item without reading.
That’s not true. The number for a feed gets bumped by one every time the feed is read and it contains new content.
My response is here
Derek, ”read” means read by the aggregator, not the human.
Wasn’t it Google who, at one time, at the bottom of it’s search results pages, gave you the option to search its direct competitors?
Yeah. As I remember, the links weren’t there if your search returned no results… Oops.
Quick links from the morning’s reading
Ringnalda talks about Google’s toolbar, Winer somehow misses his point completelyMena has moved – to a new Typepad blog!Wudi, on the other hand, now…
Dave’s Bad Mood
It would seem that Dave is in a bad mood this morning. Phil Ringnalda gave what I thought was perfectly…
phil ringnalda dot com: google and blogthis
phil ringnalda dot com: google and blogthis