Friday, August 22, 2008

Magnolia goes open source? That’s great, I guess.

image Magnolia is going open source, and will allow any site publisher to use their platform in building their own personalized version of the feature-rich social bookmarking engine. Every story I’ve read suggests the company wants to become the for the social bookmarking set.

My take: I think it’s kind of neat. I especially like the fact that bookmarks saved to user-hosted satellite installations can also be sent back to the Ma.gnolia mothership. It offers niche publishers and groups a more personalized social bookmarking option with enhanced control over their data. And, if it takes off, it could allow the company to gain market share in a field that’s dominated by Delicious and otherwise crowded with a lot of pretenders.
Nevertheless, I’m skeptical about the proposition’s overall value for two reasons.

Solving a problem that may not exist

First, is there a need for this sort of thing? Yes… and no.

Yes, I do think collaborative bookmarking for groups is a nut that needs to be cracked. Employers and other groups who share similar interests could certainly use a more focused and flexible social bookmarking environment. Delicious doesn’t offer a feature for groups, and this opening has encouraged companies like Social|Median, Twine, Reddit, and Mixx to approach the problem from different angles. The efforts in the space are at such an early stage, it seems the perfect time for Ma.gnolia to offer itself as a kind of white-label alternative to Delicious.

On the other hand, the analogy to Wordpress raises some red flags. Wordpress offered publishers and companies something that was already in the Web’s DNA: Blogging. Content management. An all-in-one Web site. But the idea of a self-hosted, user maintained bookmarking service doesn’t seem to meet the same unmet need. Why would I say that? For the same reason there aren’t a ton of group blogs out there.

Lots of people will build and host a blog for themselves or their small business. Significantly fewer people will create a blog that has multiple users. It happens, but I will unscientifically suggest that group blogs hosted on Wordpress or Moveable Type are more rare than single-author blogs. Rarer still are the individuals who host their own wikis or social networks, especially with hosted solutions like PBWiki and Ning available. Finally, Ma.gnolia and Reddit offer users a self-hosted bookmarking option? “This will be great for my friends / coworkers,” says E. Adopter! “They will love it!”

As soon as you explain it to them.

And that’s the sad truth. I have 254 friends on Facebook. I follow 90 on Twitter (with 144 followers). I’ve got 24 people in my Delicious network. Mileage may vary, as they say, but what is your gut telling you? Not a lot of people need this.

Ma.gnolia isn’t all that exciting

Second, is Ma.gnolia your first choice for an open source bookmarking service? Read Write Web’s coverage of Magnolia’s open source announcement was very positive, but it ended on a sour note:
Will communities all over the web download, customize and participate in a federated Ma.gnolia? Maybe. It's hard to know. Unfortunately, Ma.gnolia founder Larry Halff's presentation announcing the open sourcing of Ma.gnolia here at Gnomedex illustrates the problems the company will continue to face. Just like the service Halff created, the man himself seems like a brilliant guy who you know has great ideas but communicates them poorly enough that it frustrates people pretty quickly. The value proposition is unclear, the site architecture is frustrating - right now it's a service for standards true believers. This author uses it personally, though almost every time I do I grumble and ask whether I should go back to using Delicious.That’s been my experience with Ma.gnolia, too. It looks great on paper. It’s got tons of features. It’s a model for all sites that want to embrace open standards. It’s…still not a viable alternative to Delicious.
I touched on’s dominance in earlier post, but it’s worth repeating. The Delicious API makes it super easy to take your bookmarks to another service that has more features (e.g., Diigo and Ma.gnolia). Indeed, both those services allow you to seamlessly populate their service with your Delicious bookmarks using just a few clicks. And still, despite lagging in social-ness and open-ness, Delicious continues to dominate. Why? Because Delicious is better. Users can easily try out alternatives, but they keep coming back.

Ma.gnolia going open-source means you can start your own inferior bookmarking service. For people who’ve been wanting to build and maintain their own social bookmarking site, I guess that’s all they’ve got. But I can’t help thinking they’d rather be running Delicious on their servers.


Despite my skepticism, I do think an open source Ma.gnolia will be a great option for companies, groups, and classrooms that want their own little slice of social bookmarking. An open source Ma.gnolia will be especially compelling if it works on corporate and university intranets. It will also be interesting to see if the Ma.gnolia architecture will be extensible and themeable the way Wordpress is. That would be unequivocally cool.

Nevertheless, I believe that individuals who want to share bookmarks with like-minded people are better served using an existing social bookmarking hub that supports groups, which I think is just about every other bookmarking site but Delicious. The network effects, such as they are, will be more powerful than on a self-hosted site.

The fact that Delicious remains the most popular bookmarking service, even without a group feature, should tell you one of two things:
  1. No one has done groups in the right way yet. The service that perfects group bookmarking will fill a much-needed niche.
  2. Hardly anyone cares about group bookmarking.
I really hope it isn’t number two.

Sunday, August 17, 2008

Diigo loses me again.

Image representing Diigo as depicted in CrunchBase

I’ve tried to get into Diigo, the much ballyhooed Delicious alternative, a couple of times. Each time I do, I’m confronted with shortcomings that kill the deal.

First, they were even later than Delicious in releasing an add-on that was compatible with Firefox 3. Obviously, I wasn’t going to use IE while waiting for Diigo to get its act together.

Next, there were problems with its ability cross-post to Delicious. It was always a crapshoot whether the bookmarks were actually going to get there, and the Delicious update has made things even more unpredictable. Obviously, this feature isn’t a must have. Indeed, having such a feature is a plus for Diigo, because it lets users hedge their bets when deciding whether Diigo is for them. That’s exactly I’ve been trying to do, and over the last two days crossposting has been working fine, despite Diigo informing me that my Delicious account is unvalidated.

Unfortunately, Diigo’s tool-bar / add-on hasn’t worked for me all that well in my most recent testing, and I’m almost ready to give up again. Sure, it’s now compatible with Firefox 3. And it’s beautifully designed to take full advantage of Diigo’s admittedly copious social features. But when I search for items in the sidebar or on the site, the search feature is either glacially slow or it simply FAILS. As you can see in the screencap below, a search request at the site gets me plenty of ads, but no bookmarks.


If you can’t see it, Diigo’s error message reads:

We are in the process of rebuilding the tag search index as part of a major server and database upgrade to provide you with speedier performance and to support Diigo's rapidly growing user base. This may take up to 24 hours. Thank you for your understanding.

Obviously Diigo is struggling to remedy their historically slow and buggy search capabilities. But because the Delicious Bookmarks add-on handles such searches effortlessly, it’s particularly difficult to tolerate Diigo’s poor performance. Diigo may be spectacularly social, but if Diigo fails to recall my saved items, it’s failing at bookmarking. This post offers a compelling list of “six reason Diigo is better than Delicious.” But none of that matters if you can’t find my stuff when I ask for it. None of that matters if you aren’t actually bookmarking anything for me.

Maybe my insistence on being able to search my bookmarks is a bit myopic. Perhaps I should view Diigo as a purely social bookmarking service, in which reference and recall aren’t really the point. Instead, maybe Diigo is more like an advanced version of Digg or Stumbleupon where sharing is the point and personal reference is an afterthought.

My conception of social bookmarking gives equal weight to both concepts. To be a social bookmarking service your service must embrace two functions:

  1. Quick reference to my bookmarked items, with the ability to privately store items as I choose.
  2. Flexible sharing options, including both passive sharing with a network of friends and the ability to export or post items to a blog or aggregator (e.g., FriendFeed, Swurl, Second|Brain, Socialthing).

But two recent posts I read call into question my assumptions about Social Bookmarking. First, Digital Inspiration shared some data indicating that Facebook and MySpace were among the leading social bookmarking sites, well ahead of Delicious. There’s an obvious caveat to the data. As RWW pointed out, most Delicious users probably share their items with the Delicious bookmarklet—not the “Share This” plug-in from which the article’s data was pulled. But I think the posts by RWW and Digital Inspiration raise a more important question:

Since when are Myspace and Facebook considered bookmarking sites?

Neither provides easy access to your history of shared items. Certainly there’s not a reference system for your bookmarks that’s even as robust as what Mixx or Digg provide. Delicious is a bookmarking site. Facebook, Myspace, Digg, and Mixx are something else entirely, because they focus on sharing and couldn’t care less about performing the core function inherent in the word “bookmark”: I want to save something for later, for my personal reference.

Social bookmarking is about sharing. But it’s also about extending the personal reference functions that were first featured in the pioneering Netscape browser. When it comes to Diigo, my question is, “how concerned is Diigo about personal reference?” Hutch Carpenter helped me focus on this question a few days ago.

Carpenter thinks Diigo isn’t about bookmarking for personal reference, but about using your shared content as a springboard for social interaction. Carpenter’s post is a superb overview of Diigo’s social features, and he perfectly sums up the differing social concepts of the two services:

The new Delicious continues its mission of organizing a massive number of user-generated bookmarks and tags. It looks cleaner, and I like the way information is presented. Information organized by an army of user librarians. “Social” in this context means your bookmarks and tags are exposed to others, and you can find related content based on what others are bookmarking and tagging. People are the basis for discovering content.

Diigo wants people to interact via common interests in content. It has a lot of social network hooks. “Social” in this context means establishing and building relationships with others. Content is the basis for finding people.

What Carpenter’s post doesn’t address is whether Diigo is an adequate substitute for Delicious’s powerful reference capabilities. In other words, is Diigo any good at bookmarking.

I really want to use Diigo. I really want to like Diigo. But I’m not going to have two freaking toolbars cluttering my browser. And Diigo isn’t going to make me a convert until it’s good at both social and bookmarking.

Friday, August 15, 2008

Here’s a look at Mozilla’s next-gen vision for bookmarking

Obviously, it’s heavily visual, but it’s interesting to note how Mozilla plans to extend the Awesome Bar concept to dissolve the differences between your passive bookmarking (i.e., your history) and the bookmarks you actively save by clicking the now-familiar star.

Bookmarking and History Concept Video from Aza Raskin on Vimeo.

Thursday, August 14, 2008

Twine is really, really slow.

imageA while back I managed to get an invite to Twine, a new, professionally oriented bookmarking service where users can focus on sharing content with people who have similar interests. Twine claims it’s “powered by semantic understanding,” and that it can learn from its users so that, the more you use it, “the better it understands your interests and the more useful it becomes.”

I think semantic-powered, interest-oriented sharing is a powerful idea. Certainly, as a lawyer, I can see how creating a community to share and discuss highly specialized knowledge could prove incredibly useful to a whole class of knowledge workers (e.g., physicians, librarians, academics, journalists). Also, I think collaborative notetaking for groups is one of those nuts that hasn’t yet been cracked. It’s a problem that needs solving because I haven’t been impressed with the way traditional bookmarking services, such as Ma.gnolia, have handled their group features.

And Twine is pretty. It’s well thought out, and its features are smartly scoped. But I’ve been holding off reviewing the service because of one HUGE drawback. Twine is really, really slooooow. It takes a long time for each page to load (rough estimate: forever). So I’ve barely been able to use it. Responding to a recent comment in Twine’s Beta Feedback Group, one of the developers admitted there was an issue:

There are some issues on the server side that we are working on (such implementing more caching). There should be significant improvements in the next release (we're hoping to do a patch this week or next week at the latest). There will be further improvements on an ongoing basis for the next 2 months. Remember it's in beta and we're still working on it, so thanks for your patience. I think after the patch it should be faster. Let us know.

Also if you are using Twine over international lines or lower bandwidth connections that could also be affecting performance on your end.

A few ways to speed things up a bit -- go to your Account page, and then select Twine Settings and select only the twines you really care about to show up in your Interest Feed and the Bookmarklet. That should speed things up a big for you as well.

Twine isn’t one of those services you can just figure out immediately, so I’d like to make more rigorous use of the application. But until these speed issues are figured out, there’s not much I can do.

Wednesday, August 13, 2008

Tumblr might be fumbling its new release

Andy Desoto is disappointed that Tumblr has created an entirely new beta site just to try out a larger sign-in box.


I thought Tumblr was fairly clear in their post about the new site. They’ve redesigned the dashboard in preparation for a much bigger overhaul:

We’re getting ready to push our biggest interface overhaul since last November, and we’d love to get your feedback.

Nearly every piece of the interface has changed. We’ve tried to perfect the presentation of the Dashboard, and the retooled interface will enable a lot of the new functionality we’re prepping to launch.

In other words, the update hasn’t happened yet. But if people are confused on this point, it’s Tumblr’s fault. The Tumblr Beta site doesn’t look very different, and there’s nothing inside to indicate new features might be on the way. It might have been better to wait until the new features were ready before unveiling the public test site. Indeed, that’s what a test site is for: creating new features and functionality and working out the bugs so that you can avoid confusing your mainstream users.

So far, confusion isn’t being avoided.

Why create a custom search engine out of your bookmarks?

I keep seeing people share this RWW post, in which Sarah Perez explains how to make your social bookmarks into a custom search engine. I can’t figure out why people think it’s such a great idea. Sure, a simple text search of your bookmarks is lots faster than simply browsing your tags. But Perez’s idea involves:

  1. Exporting your bookmarks from your current bookmarking service.
  2. Importing them back (using Posterous) so that they’re contained on a single web page.
  3. Adding that page to a Google Custom Search Engine.

Oh, and after you’ve done all that you still have to use a bookmarklet if you want to add new items to your search engine.

As people in the comments to the RWW post make clear, there are much easier ways of searching your content:

  • Use your existing bookmarking service. For example, the recently upgraded Delicious now offers even faster search, using both text and tags. And the Delicious Bookmarks plug-in has always been speedy when it comes to searching your bookmarks.
  • Use one of the newish content aggregators to search your stuff. If you’re going to use a different engine, why not use FriendFeed, which can search all your items (not just social bookmarks)? Second|Brain is even better, since it creates a database of your private items and keeps them private. I criticized Second|Brain in this post, but the recent update fixed a lot of the problems I was having and added a ton of new imported services.
  • Go passive. If you truly believe social bookmarking has nothing to offer you any longer, then forget about actively saving anything at all. Just search your Web history. As I point out here, if you opt into Google Tool Bar’s Web History feature, all your Web searches and activity are silently recorded and you, can search this humongous database any time you want. Google Web History also searches your Google Bookmarks. Privacy-niks may balk, but using Google’s custom search engine and Posterous certainly isn’t any more private. If you want privacy, try Firefox’s new Awesome Bar. It searches both your bookmarks and Web history, as well. All you have to do is start typing in the address bar. Hard to get easier than that.
  • Use Google. The easiest thing is to just search for stuff in a traditional search engine. A few responses to Mathew Ingram’s “Who bookmarks any more?” post suggested exactly that . Sure the database is many times larger and the results might not be 100% relevant, but at least your’re not bookmarking anything or creating more work for yourself.

The bottom line is this: You either like social bookmarking stuff you don’t. If you don’t think there’s value in bookmarking, then just consult your browser’s Web history or use Google. It’s faster and easier. But if you do like social bookmarking, there’s absolutely no reason to re-invent the wheel. You can more easily search your bookmarks using existing tools.

Monday, August 11, 2008

Social|Median was not a good experience

image After reading Louis Gray’s positive post the other day, I tried out Social|Median. Perhaps I wanted the wrong things from the service, but it didn’t work all that well for me. Here’s what I thought I could do:

  • Start my own mini-Digg news network focused on a single keyword (in my case, “bookmarking”) while also selecting some additional, secondary keywords (e.g., social media, technology). Items matching keywords would automatically be added to news river that could be sorted either by popularity or date.
  • Add my own items via a bookmarklet. I could choose either to add items directly or let Social|Median choose.
  • Import RSS feeds, including my own and others that matched my content preferences.
  • Adjust the “noise meter” so that items with less relevance to my keywords (as I ranked them) would be filtered out.

Because I consume news at a much faster rate than I write it, I had hoped that Social|Median would be an interesting place for me to collect items related to this blog’s focus: web-based bookmarking and notebooking. Unfortunately, when I tested the service, it did none of the things I thought it would do:

  • After adding a patch of items immediately, including items from the feeds I had chosen, it was dormant for several days.
  • When I tried to add items to my network via the bookmarklet, they wouldn’t actually show up in my network. This was true whether I let Social|Median choose or whether I added items manually.
  • Although I tried to import a few feeds from elsewhere, none of those items were added to the SM network.
  • Adjusting the noise meter had no discernible effect.

In fairness, other users are probably using Social|Median to consume a broader spectrum of news. I was trying to collect news within a fairly narrow focus, and I tried to start small—importing only a few feeds in addition to my own. Here’s the network I created. Notice that a few items have since been added, but far fewer than I would expect given the fairly broad keyword “bookmarking” and the other feeds I’d selected.

Again, it could be that I misunderstood the capabilities of the service. It could also be that there are some bugs that need to be worked out. I’ll never know: I submitted a support request mentioning the problems, but it went unanswered. Then I suspended my account. I was presented with a form asking why I was deactivating my account. I explained the problems again in the fields provided.

I’ve received no response. Which is fine. They’re busy.

But so am I.

Social|Median: F

Saturday, August 9, 2008

Tumblr is an image bookmarking service and (potentially) a whole lot more

Last week, after reading Corvida’s RWW post on image bookmarking services, I started to wonder why I’d never felt the need for such a service. I’m obviously a fan of social bookmarking, and my preferred service, Delicious, has no image functionality. Even before Corvida’s post, I’d been aware that Ffffound! and could fill the gap; I just hadn’t signed up. I've just been sorta “meh” about image bookmarking.

Then I realized why: I’m already using an image bookmarking service. It’s called Tumblr. I use Tumblr’s amazing bookmarklet to save images all the time to my tumblelog, It’s true no one thinks of Tumblr as a bookmarking site. It’s certainly a popular microblogging service, but it doesn’t have the obvious social components inherent in a bookmarking site. Which is to say it’s not an unfathomable zoo of OPC (other people’s content).

Tumblr’s image bookmarking bona fides

But let’s talk about what it does. Its bookmarklet saves a new copy of any photo to the hosted Tumblr site; you’re not hot-linking to some other site. And the process could not be easier:


I can save photos, add tags, and add unlimited text comments, including HTML. And it shows up perfectly formatted on my own Web site.


But what if I want to find it later? Tumblr doesn’t yet have built-in search, and its tagging features are admittedly embryonic. But, check it out! Tumblr has visual archive that’s much more useful if I’m looking for a specific photo. Note that I don’t have to remember what tag or text I used to describe it; I can search over months of archives in mere seconds and spot the thumbnail I need. For good measure, each thumbnail does include a good-size portion of text, which is searchable using the browser (ctrl-F).


What Tumblr doesn’t have

Tumblr is, first and foremost, a powerful blogging platform and content aggregator. I don’t just use it to collect images. I can post videos, quotes, conversations, and mp3s, all using the same flexible interface. I can also import feeds from Google Reader, Delicious, Twitter, or any other service that makes RSS feeds available. But as much as I love Tumblr, it’s NOT a true substitute for a social bookmarking service.

Essentially, Tumblr = social bookmarking. The strength of any social bookmarking service is the recommendation engine—quickly discovering what others have found interesting. Or beautiful. Or scandalous. Or sinfully, horribly cute. Tumblr doesn’t have a recommendation engine.

Not yet, anyway.

But the potential is there. Tumblr already has the guts of a great social bookmarking service. Its got tons of active users submitting content in a fairly standardized manner, either through the bookmarklet or by importing feeds. It also has a full-featured, read/write API. There’s really no reason all that data shouldn’t be leveraged into a Tumblrmeme or Tumblr-icious.

Tumblr has taken baby steps in that direction with a number of features. For example, the Tumblr Radar showcases content from around the Tumblr-sphere. Like real-life radar, stuff tends to fall of without a trace, and there’s no RSS feed, but you can browse recently featured Radar items here.


Every user’s dashboard features a Livejournal-like stream of activity from the Tumblr blogs you’re following. You can choose to re-blog items that your friends have blogged, and popular items feature a handy counter showing how many times an item has been re-blogged. Unfortunately, I don’t keep up with my dashboard as much as I’d like to. That’s because (wait for it…) there’s no RSS feed.


So yeah, unlike Delicious or Stumbleupon or FriendFeed, Tumblr isn’t exactly optimized for content discovery. There’s no way to search for items or sort by tag (although Yahoo! Pipes ninja Joe Lazarus provides a nifty tracking option). And you can’t easily identify popular items, because Tumblr doesn’t have a robust way to measure the zeitgeist. How cool would it be to see Tumblr fed into popurls? The Radar and Dashboard features prove it’s not unthinkable. Indeed, with Tumblr promising a major upgrade soon, I’m hoping to see some progress on this front.

Even if Tumblr doesn’t create an enhanced portal for exploring all that content, the API seems like it would allow another developer to do so. But Tumblr hasn’t really caught fire amongst the tech bloggers and developers. Tumblr has been criticized for its notoriously long dev cycles and the stingy pace at which they add features. Rather than adding features every few weeks, it seems they’d rather unveil a complete overhaul after a long period of dormancy. That’s a great strategy for an occasional big bang, but it’s not a good way to keep bloggers and developers interested in your product.

What could be…

Tumblr may not ever generate the breathless kind of buzz FriendFeed and Twitter have enjoyed. For me personally, Tumblr has been (by far) the more useful and rewarding product. It gives me a little place the Web that doesn’t look like a hodge podge of voices; it’s my voice. But underneath this flexible and intensely personal blogging platform is a sturdy framework that could support a world-class bookmarking site and recommendation engine. It’s these dual possibilities that make Tumblr so fascinating.

Tumblr founder David Karp is aiming for a million users by the end of 2008. Imagine peeking inside a million separate hearts. At once.


UPDATE: I should point out that the aforementioned Joe Lazarus has built a few good Tumblr aggregators including one for frequently Tumblr'd photos, popular quotes, and one for songs. None is the all-in-one portal I'm thinking of, but he gets bonus points for actually using Tumblr to create his mashups.