An updated version of the prohibition on burning the candle at both ends

Here’s why it’s important to test plugin updates.

After my last post about beta testers for WPBook, I decided to go update my other plugins which had updates available, including Lifestream, which had an update to 0.99.9.8-BETA from 0.99.6 available.

So I jumped in without really doing any investigating of what changes there were – bad idea.

Here’s what I got for my mistakes:

Lifestream Gone Wild

Somewhere between whatever version I was running (I believe it was 0.99.6) and this current 0.99.9.8-BETA, the Lifestream developers changed the way they track new events, and started to “publish” every Lifestream event as a post, using custom post types as defined by WordPress 3.0. Unfortunately this wasn’t stated very clearly in the documentation.

I’ve deactivated the plugin and deleted all the extraneous wall posts Lifestream created – hopefully not too many got passed into my friends streams.

I’ll have to look at how WPBook can better handle “custom post types” and perhaps create a setting whereby folks using custom post types can decide which post types WPBook should and should not cross post to Facebook.

I think ultimately I’ll want to keep an open source (LAMP) platform because I want to own the data in my lifestream, have backups of it, and be able to move it around as I please. This leaves me choosing between a platform linked to a blog (WordPress or MovableType) or a standalone one (Sweetcron, Storytlr or similar) that just powers the lifestream. Originally I created JohnEckman.com as a standalone lifestream, thinking that the various blogs I wrote for around the web could be aggregated there – but there’s no reason why that couldn’t be a WordPress install as well.

Anyway, what follows are my notes / first impressions – not an exhaustive evaluation certainly but a good shortlist to start with if you’re thinking of running a lifestream.

First, the open source platforms

WordPress with the Lifestream plugin by David Kramer from Enthropia Labs. You can see this one in action on Open Parenthesis.

WordPress Lifestream Plugin Output from Open Parenthesis

I really like the plugin’s approach, which is to leverage WordPress and enable you to put your lifestream up as a page, as well as providing a widget you can put on your blog homepage. (Of course you can use WordPress without using the blog engine at all – in which case it is more like the standalone options). It’s very simple to install and configure, and supports a wide variety of places from which you might want to pull feeds:

Lifestream Plugin for WordPress Add Feed Screen

The Lifestream plugin can group items by day to avoid clutter (days when I work from home and listen to lots of tracks on Last.fm, or upload 30 photos to flickr, that single source can easily overwhelm a lifestream) – though that isn’t available for all feed types. The plugin is also themable and extendable, which is key – you can add feed types, custom icons, or change display options at will.

Movable Type Open Source edition with the ActionStreams plugin. I was very excited when the Action Streams plugin came out, and set up JohnEckman.com on MT just to be able to run it. Heck, I even wrote a few plugins for Action Stream parsing. However, I’ve been really lax about staying up to date with new releases. Having a number of blogs already on WordPress makes it far more likely I will keep that updated. I know many folks use and love Movable Type but its mental model of how a blog works is just not in synch with mine the way WordPress is.

MTOS based Lifestream from JohnEckman.com

That said, if you’re used to Movable Type’s approach, using the Action Streams plugin certainly provides a flexible, pluggable, themable way to aggregate your lifestream.

Sweetcron an open source application built using the Code Igniter framework – see this one in action at nastyhack.org/sweetcron/.

Sweetcron based Lifestream

Unfortunately, the original developer has since moved on to other things and no replacement has clearly emerged yet. I like that it operates as a kind of standalone platform, not tied to a blogging engine, and gives you good theming potential – but I worry about the long term prospects for the project without a clear leader organizing and directing contributors. On the plus side, there are many great themes for Sweetcron, and if you’re familiar with code igniter as a framework .

Storytlr – what was once a SaaS option but has ceased operations and released their code as an open source project – you can see my lifestream at nastyhack.org/storytlr/.

Storytlr based lifestream

There will be some work to do make the project more amenable to self-hosting (it was built to operate as a service), but it has the benefit of starting from a solid working foundation, not starting from scratch.

They’ve also got a multiuser version, something which was often requested for Sweetcron but wasn’t forthcoming. (Will the WordPress Lifestream plugin work on WordPress Multi-user, which is being merged into the core code in 3.0? Don’t know yet).

Hosted offerings:

Chi.mp My Account) – Hosted service which serves as an OpenID provider as well as aggregating your lifestream. Integrates with Twitter, Facebook, Flickr, any RSS feed. Also enables you to import contacts from online address books. Seems to have gotten stuck on some old photos in its Flickr import and is failing to import newer content? (Haven’t spent much time troubleshooting here – mostly using it as a backup OpenID for when my primary one fails). Their focus has always been on enabling you to own your own data – allowing you to export all of your contacts and updates out of the site as a zip, which they still do, with appropriately microformatted content throughout. If I were to rely on a hosted version, this would be my top pick.

Flavors.me (My Account – but check their user directory for better examples) – Hosted service, supports (as of 12/24) flickr, tumblr, twitter, vimeo, last.fm, facebook, goodreads, netflix, and generic RSS. Gets bonus points for using OAuth appropriately and not requesting usernames/passwords where they aren’t necessary. Offers some basic layouts and admin-side design flexibility (fonts, colors, backgrounds, and the like. (See this great review on the Lifestream blog for more info).

FriendFeed (My Account) – Arguably the giant in this space until getting acquired by Facebook. Probably keep this around as long as it exists, and do sometimes follow folks here, but not somewhere I’m spending much time anymore.

Plaxo Pulse (My Account). Plaxo seems to be where my less internet-savvy family and friends get updates. Folks who aren’t on Facebook, or Twitter, or spending lots of time reading blogs and other RSS feeds, but get a periodic “pulse” from Plaxo of what’s going on in their network. For that reason alone I’ll probably never take it down. They’ve also got Joseph Smarr and John McCrea, who’ve been pushing for the open social web, open standards, and data portability. (Updated: Smarr’s moving to Google). They’re supporting OpenID, OAuth, Portable Contacts, and other new standards as soon as they are introduced, and I like that they’re good web citizens (and have continued to be, post acquisition by Comcast). That said, it still feels to me like a downstream destination for a specific kind of consumer, not the central place I’d think of as my lifestream.

Ones I haven’t set up or tried to use: Lifestream.fm, AIM Lifestream.

What are you using? What have I missed out on altogether?

It’s a very vibrant space right now, with commercial vendors, open source projects, trends, and standards all vying for attention. I’m thinking here of a couple of overlapping categories:

OpenID

A standard, for which there are good open source libraries, but also commercial providers. Increasingly I’m seeing OpenID as one service (often the anchor service) provided as part of a suite. Of course the traditional mainstream web players like LiveJournal, WordPress.com, Yahoo! and AOL (through AIM) are providing OpenIDs as well.

Lifestream and Profile Aggregation

More social networks == more profiles, and more feeds. A number of services/projects have sprung up (I talk about a few below, but there are many others as well) which enable you to aggregate together all of your profiles in a single place. Some are more focused on aggregating all of your feeds – creating your lifestream and letting others subscribe to it; others are focused on aggregating the feeds of your friends, to make it easier for you to follow.

The Data Portability Project, Open Web Foundation, Open Social Foundation, Autonomo.us

These foundations are not focused (directly) on producing software, but on building awareness of and consensus about the need for user freedom on the internet, and publishing open specifications which will lead to a world in which our online identities and data streams can be more easily managed, exchanged, and even migrated from provider to provider.

It would really be a full-time job to keep track of all that is going on in this space, but here are a few I’ve been following / trying out.

Chi.mp

Chi.mp enables each user to create their own domain in the .mp TLD space. You can check out mine at johneckman.mp. Chi.mp provides OpenID, but doesn’t (yet) consume it. (I can use johneckman.mp as an OpenID to log in to other sites, but I can’t login to chi.mp with an OpenID from elsewhere).

Chi.mp also supports a number of services (currently Twitter, Facebook, Flickr, Gmail, Yahoo (mail), and Hotmail). For those which provide activity feeds, chi.mp will pull those feeds into your profile (viewable by others) and dashboard (viewable only by you). There’s also a generic RSS feed import capability, for services (like personal blogs) that Chi.mp doesn’t know.

Finally, Chi.mp also supports your social graph – your friends lists or contacts lists from various services can be imported – from webmail services like gmail and hotmail but also from services like Twitter, Facebook, and Flickr which have contacts or friends exposed via an API.

It’s very highly configurable, in terms of who can see what. You can tag contacts, and tag feeds, and use tags to determine visibility of feeds to groups of contacts. I haven’t yet really figured out what else Chi.mp will be able to do with webmail services – I don’t think I will ever want emails I send or receive showing up in my action stream or on my profile, but certainly being able to leverage various APIs for getting contacts will reduce the need to “refriend” people on each new network.

For now, however, Chi.mp has no way to identify that the “John Doe” you are friends with on facebook is the same person as the “John Doe” who is a contact on Flickr – they provide a simple way to manage contacts (and “merge” the two contacts into one virtual person) but there is still human effort (decision making) involved in reconciling these graphs.

Trufina

Trufina adds an interesting twist in that they are trying to more tightly link online identity to offline. Using a method well known to financial services companies – the ability to answer a short set of questions about your financial history which would not be known to someone who found your wallet in the street – Trufina verifies that the person using the name John Eckman is the same one who lives at a given address and has other “meat space” attributes.

You can see my default public profile here:

I have to say their focus on “criminal background checks” I found a bit creepy: I realize that background checks are important for certain kinds of employment, but it seems like the need (and even desire) to assert a record free of felony convictions should be a niche market, not the default market for an online identity vendor. (Employment verification and educational background verification are said to be in development).

They also then enable you as a user to share various parts of your verified identity with others, including inside a number of social networks. You can create an “ID Card” and show that to only specific folks. There was no way I could find, however, to not show the “Criminal Records Search” section of the ID Card – it seems to always show either “” or “ ” – neither of which makes for high confidence. How about the ability to not show that section at all, if I don’t think it’s relevant, in which case you could also dispense with the rather elaborate disclaimer about criminal record checks.

I also found it frustrating that the default “profile” view – the only one people can get to who haven’t been specifically authorized by you – shows only the Trufina user name. Obviously given the market Trufina is after, and the data that a full profile might ultimately contain, they need to be concerned about privacy. But what if I’m perfectly happy to have people see my first and last name and maybe state of residence and employment?

The key to privacy needs to be control, not defaults which prevent users from making basic data public.

Identity.net

The folks at Identity.net have partnered with Trufina, to link your “Trufina Verified Identity” to an OpenID which can be used throughout the web. It’s a great concept – to be able to demonstrate that the virtual identity a given OpenID represents is tied to a real offline person could be quite valuable.

I’d hope to preserve, however, the option to also have OpenIDs which are not linked to my offline identity. The ability to get the benefits of OpenID (in terms of single-sign-on) without necessarily having all online activity tracked directly to your offline identity is one of the freedoms the internet promises and I’d hate to lose that. (A number of OpenID providers enable you to create multiple OpenIDs that only they know are associated with each other – this enables you to project different identities on different sites).

Like many of these services, Identity.net is in beta, and was having trouble with their control panel when I signed up, so it’s possible I haven’t yet seen what flexibility they offer in creating and using OpenIDs tied to your Trufina identity.

ClaimID.com

I use ClaimID with delegation to use JohnEckman.com as an OpenID. ClaimID also provides a basic profile page on which you can enter links to web sites and verify your ownership of them, as well as display contacts, optionally marked up in XFN (with semantic data about the relationship you have with each contact).

Movable Type with Action Streams Plugin

I use the open source edition of Movable Type with the Action Streams plugin to power JohnEckman.com as a lifestream aggregator, pulling in feeds from various web services. Creating additional plugins to add services to Action Streams is relatively simple, and I’m hosting it myself so I have complete access to the data stored and complete flexibility in display.

The DiSO project has produced a similar plugin (WP-DiSo-ActionStream) for pulling action streams into WordPress (I’m using the OpenID plugin from DiSO on this blog), and there’s an Activity Streams module available for Drupal as well.

Finally, there’s Sweetcron, which I have only just started to experiment with, but which focuses on just managing the action stream aggregation without the extra functionality (and overhead) of a blog or other framework like MT, WordPress, or Drupal. It’s also easily extended.

As all of these evolve, how much ownership and control will users want to have over the content their online activity produces? How much technical understand and effort will they be willing to expend in order to exert that control?

What, in other words, will be the balance between hosted providers (they do all the work but also retain some element of control) and self-hosted open source platforms (you do more work and gain more control)?

Will the central difference between the two options lessen as real data portability becomes commonplace?

The problem is that when I send brightkite a message like: @bos

It sometimes knows that is Logan Airport in Boston, and sometimes just knows that as BOS, United States, which results in updates like this:

Anyone know why this is happening? Brightkite’s fault? Fire Eagle’s fault? Neither?

Update: Should perhaps have checked first. Brightkite seems to be working on this issue, at least according to this post on GetSatisfaction. Seems like they recognize some airport shortcodes but may not always have the full name associated with it and what city/state they are in.

Since my FireEagle feed only updates with city level info, you can’t see that it is an airport.

Hopefully this will mean fewer “United States” updates in my feed over time.

TripIt has launched profiles for travelers, with some pretty good controls on what is public and what is private:

The immediate goal is to give TripIt travelers one place to track all their travel information and showcase their travel history. The profile includes basic information about a traveler, including home location, upcoming trip destinations, connections in TripIt as well as important travel statistics like miles traveled, days on the road, etc.

It’s got a nice, RESTful public url – mine’s at http://www.tripit.com/people/jeckman

I’ve updated my TripIt Action Stream plugin – the good news is that it will now provide a real profile link rather than just linking to the TripIt homepage.

You will, however, have to make your activity feed available to everyone – but if you didn’t want to do that, you probably don’t want to publish your activity feed as an action stream anyway. (Actually you could leave your activity stream private and still publish your profile link – just uncheck the activity feed checkbox when adding the profile inside MT).

It’s basically a roll-your-own lifestream, though for now at least it isn’t integrated to anyone else’s streams, as in Friendfeed or Socialthing.

This morning I posted a new plugin which picks up TripIt Activity Streams.

TripIt’s activity stream is a private Atom feed which posts an item whenever you begin a trip, complete a trip, or start planning a trip. For example, here’s a recent entry from my feed:

You can download the plugin from the MTAS page.

Check out the work in progress at johneckman.com. Read on if you’re interested in creating your own action streams.

Although it has been a while since I’ve worked in Movable Type, it was a relatively painless install (assuming you’ve got the basic LAMP stuff in place).

The challenge was creating additional “services” for this blog and goatless. There’s some basic info on writing new services in this blog post: “building action streams” – but it took some time to translate this into my way of thinking about it, including where the png file goes for the icon for the service.

Here’s what the config.yaml looks like for the OpenParenthesis service:

name: Open Parenthesis
id: openparenthesis
key: openparenthesis
author_link: http://johneckman.com/
author_name: John Eckman
description: Adds profile service and activity stream for Open Parenthesis
version: 1.0
plugin_link: http://johneckman.com/

profile_services:
    openparenthesis:
        name: Open Parenthesis
        url: http://www.openparenthesis.org/
        icon: images/openparenthesis.png

action_streams:
    openparenthesis:
        posted:
            name: Posts
            description: New Posts to the Blog
            html_form: '[_1] blogged on <a
                href="http://www.openparenthesis.org/">OpenParenthesis</a>
                about <a href="[_2]">[_3]</a>'
            html_params:
                - url
                - title
            url: 'http://feeds.feedburner.com/OpenParenthesis/'
            identifier: url
            xpath:
                foreach: //item
                get:
                    created_on: pubDate/child::text()
                    url: link/child::text()
                    title: title/child::text()

This config.yaml file goes in mt/plugins/openparenthesis/config.yaml – that is, you create a new plugin folder containing just this config.yaml.

Then you put the openparenthesis.png in: mt/mt-static/plugins/openparenthesis/images/openparenthesis.png

I suppose much of this would have been obvious had I been more accustomed to how MT plugins work.

You may also note I’m just hardcoding a pointer to my feedburner feed – so technically there is no “{{ident}}” needed – it doesn’t matter what identifer I put in.

One could translate this by changing:
url: 'http://feeds.feedburner.com/OpenParenthesis/'
to:
url: 'http://feeds.feedburner.com/{{ident}}/'

And then entering the feed name (the only part of a feedburner feed which changes) into the identifier field when adding the service to your profile from the MT admin side. That would give you a generic FeedBurner feed, to add blog postings to your action stream.

Of course, this may not make sense if your blog is itself on Movable Type, since your blog postings will be directly in your feed, but if you’re creating your action stream on separate domain, it may come in handy.

You can create extra “profile” links which offer no streams as well. For now I’ve just cheated with the link to Optaros at the top of my “Find Me Elsewhere” list.

Matt and co. at Automattic released Prologue, a WordPress theme (GPLv2) which creates a twitter-like experience based on posts to a WordPress blog. (It’s already been updated once).

Check out the Prologue Demo Blog for a sense of how the theme works. This could easily be used to create a kind of workgroup twitter, and given the number of different plugins / mechanisms for creating a blog post it could be extended to mobile, IM, and other integration points. The important difference, of course, is that you’d be hosting your own experience, not relying on Twitter – though that also means you’d need to build your own audience.

The folks at SixApart released the Activity Streams plugin for Movable Type which

lets you aggregate, control, and share your actions around the web as well as a list of your profiles on various services. With the Action Streams plugin you keep control over the record of your actions on the web. And of course, you also have full control over showing and hiding each of your actions. The Action Streams plugin, by default, also publishes your stream using Atom and the Microformat hAtom so that your actions aren’t trapped in any one service.

You can see a great example of this on David Recordon‘s site (he’s the Open Platform Lead for SixApart) and in a group context on the Movable Type Activity Stream page.

Both of these represent significant advances toward an open source, open standards, portable data approach to social networking and lifestreaming.

Since the implementations are open source, expect similar functionality to be ported across platforms.