skippy dot net

Google Exodus

I've been using Google Reader daily since December, 2006. I use it to read web sites that interest me. Although some people have complained that it hasn't been updated in a long time, or that it lacks meaningful social aspects, I've been perfectly happy with it.

I don't generally read the news on my mobile device. Any sharing of news I might do would likely occur manually through email or Twitter, rather than some button on the Google Reader site. The lack of new features is actually something I like: it's a product that works well, and has continued to satisfy my very modest needs.

But like all good things, Google Reader is coming to an end. There's a general scrambling amongst Reader users who are looking for alternatives, and there are a number of solid contenders out there. Most of the current crop are offering features that Google lacked, which means they're not features in which I have a strong interest. I expect I'll wait a little longer to see what options mature in the time before Reader officially shuts down.

I think my preferred solution will be to run my own RSS aggregator again. Although the format hasn't advanced substantially in the half-decade that I've been using it, I think the tools available to produce and consume RSS have matured quite a bit.

Just a year after adopting Google Reader, I went all-in with Google for Domains. I updated my DNS MX records, and handed control of all of my email to Google. I eventually switched from using a desktop email client to using Google's web interface exclusively. On the whole, since November 2007, I've been pretty happy. The Google experience is perfectly acceptable.

But the shuttering of Google Reader has me asking: "What else?" It's extremely unlikely that Google would retire their mail service, but what if? Since I'm expecting to run my own RSS aggregator in the near future, does it make any sense to reclaim any other services back from Google?

Way back when, the big draw to Google Mail -- for many, including me -- was the superb web-based experience. From any computer with a web browser, I could access my email. Self-hosted web-mail solutions like SquirrelMail, Roundcube, and Citadel all valiantly tried to compete, and I'm sure they each have a healthy following; but they didn't do it for me.

In today's world, with the proliferation of smartphones, I have convenient access to any email solution, whether mine or someone else's. Web mail is no longer a strong distinguishing factor for a self-hosted email solution. I suspect I could get by without web mail at all by using my phone. If I'm not using web mail for on-the-go access, it makes little sense to use web mail for desktop access. It's been a long time since I've used Thunderbird -- or any other desktop email client -- but I suspect the learning curve would be mild.

Over time, Google's integration of their calendar function and their chat functions made the Google Mail web interface even more useful. And of course, there's Google's superb search backing up all that email. If I were to bring email back from Google, I'd need to tackle the first of those two issues. (Search is largely a non-issue for me, as I very rarely keep my email, let alone search it.)

I've used Zimbra in the past, and could use a full-fledged collaboration suite, if I wanted to. I think things like Zimbra and Citadel and the like all scratch itches I don't have though. I mostly need email, and a decent calendar and contact manager. ownCloud offers the latter in a very nice package, so that would probably be my first stop.

It's been a long while since I last ran my own mail server, so I've got a lot of learning to do should I pursue this.

Aggregating

The value proposition of social media sites like Twitter has always been somewhat vague to me. I've stated before that I'm skeptical of social media, and that I'm not one to jump on social network bandwagons. I recently purged a bunch of people from the list that I follow on Twitter because I wasn't seeing any value to reading what they had to say. There's only so many hours in the day, and I'd prefer not to spend them reading about what other people had for lunch.

I know that part of my problem with aggregating too much information is the workflow I use. I'm extremely linear when I process things: I work from oldest to newest when reading news in Google Reader. It's only in the last couple of months that I've started marking whole categories as read, even if I hadn't read them: "if I'm not reading them, why am I aggregating them?" is the question I ask myself. When I reload the Twitter home page, I scroll down to the last thing I read (or the bottom of the page, if I'm that far behind) and then work my way up. I rarely page back to see items pushed off the home page. I use the Twitter home page because I haven't found a dedicated Twitter client I like.

But the thing that's really stuck in my craw right now is duplication of information. Most of the people I follow on Twitter are also people included in my list of feeds in Google Reader. Whenever someone posts a new blog entry, there's almost always a Twitter message declaring that fact (our software automates this for us). I almost never click the link from the Twitter message to the blog post, knowing that the post will eventually be picked up by Google Reader for me to review. Most of the people I follow on Twitter also tweet enough other stuff to make it worth continuing to follow them on that service. A notable exception is, interestingly, CrunchGear: the overwhelming bulk of the CrunchGear tweets are simply the new posts that have gone online. Since I'm aggregating CrunchGear in Google Reader anyway, what's the value in following them on Twitter?

I could, of course, aggregate the Twitter feed(s), so that Google Reader is my sole source of incoming information. But I've noticed a pretty big lag in Google Reader most days, such that a tweet posted early in the morning by someone might not be displayed to me in Google Reader until mid-afternoon. Most of the time, this might not be a big deal, but every now and again someone will tweet something that merits an immediate response: either a question for which I know the answer, or a request for a recommendation, or even an invitation. These things can be time sensitive, and I'll have missed the window of opportunity if I rely on Google Reader catching them and displaying them to me.

It's this delay that also prevents me from using something like Yahoo Pipes to create some kind of filter to weed out the extraneous bits, so that I can focus on the compelling data from each disparate service I use.

The thought that started this little tirade was the idea that I might integrate my Twitter posts directly into my blog, in a fashion similar to Chris' lifestream. Rather than a dedicated page, though, I would simply grab my tweets and store them as a new Habari content type for display alongside my normal posts. I could then also include my Flickr photos, and whatever else I wanted, making the front page of my site the complete clearinghouse for all my online activities. Then folks could simply aggregate one site to follow what I'm doing.

It's a nice idea, but it fails in execution. In addition to the delays noted above for feed readers acquiring new data, the convenience of replying on Twitter is made more complex: a reader would have to see in my feed what I had posted to Twitter, then go compose their reply either at the Twitter site or in their Twitter client. Similarly for commenting on my blog, or on any Flickr photos I posted: following the lifestream is just one piece of the puzzle. Interacting with the information presented in that stream is the next hurdle.

What do you think? How would you like to simplify and integrate interactions with aggregated information?

Google for Domains

I've been looking at outsourcing my email, and after writing that post I signed up for the Google for Domains service. What follows are my initial thoughts now that I've started using it, as well as some suggestions based on my experiences thus far (mostly aimed toward my family, who are now also using Google for Domains).

The Google Control Panel is extremely easy to use. It took me all of four minutes to get things set up. Google makes it easy for you to use links to your own domain in order to access their services, and they even provide to you the necessary DNS changes in order to support those custom URLs. So instead of http://mail.google.com/a/example.com, you can use http://mail.example.com. The same goes for docs.example.com and calendar.example.com. Adding and managing users is similarly easy.

I selected Google for Domains primarily for their email service, so that I can stop managing my own MTA and anti-spam tools, so that was the bit in which I was the most interested. I host a handful of domains for my family, and I wanted to make sure all their vanity email accounts remained. Thankfully, the Google control panel allows you to associate additional domains to the primary domain, and again they provide instructions for setting up the DNS MX records as needed. The one catch to this is that an email address is valid across all the domains you define. So if I define domains example.com and example2.com, email sent to user1@example.com and user1@example2.com will be delivered to the same user. This isn't a big deal for me, but it might be a gotcha for others.

I've been accessing my email for the last several years by IMAP, using Mozilla Thunderbird. I have a handful of folders into which I organize my mail. Not having spent much time with GMail in the past, I was a little worried about workflow changes that might be required of me. In particular, I was worried that all the thousands of messages I've been saving over the years would be stored in one gigantic Inbox, and I'd see the gigantic list of messages every time I logged in. Thankfully this was not the case, and importing all of my old mail into Google Mail was an easy (if extremely slow) process.

Google's web-based mail system does not use folders like Thunderbird; instead it uses labels. Labels are just that: it's a term you apply to a message. You can then filter the list of messages you see in the web-based interface by label. Messages can have more than one label, which provides some additional organization convenience over folders, since a message can only be stored in a single folder at a time (unless you copy it into multiple folders, but then deleting the message from one folder will not automatically delete it from other folders into which it might have been copied).

By default, messages do stay in the Inbox, even after you label them. It's easy to move old messages from the Inbox, though, through the use of the Archive button. When a message is marked "archive" it is removed from the Inbox, but remains accessible through search and label filtering, as well as the special "All Mail" display. In essence, labeling a message and then pressing "Archive" is the same as moving a message into an IMAP folder.

Google Mail does provide IMAP access, so you can use a traditional mail client if you prefer. Importing messages requires that you connect via IMAP. When accessing your mail via IMAP, the labels you created are presented as IMAP folders. If you place messages into the Google Inbox, they will be visible in the Inbox when using the web-based interface, as well. If you place messages directly into a Google label-as-IMAP-folder, it will acquire the appropriate label and be automatically archived, preventing it from being displayed in the Inbox. So the process of importing all my old mail was simple: define a new IMAP account in Thunderbird for Google Mail and then copy messages from my old IMAP folders into the corresponding Google label-as-IMAP-folder. I elected to create the labels in the Google web interface ahead of time, for no particular reason. I assume you can create new labels-as-IMAP-folders using Thunderbird and have that work the same as if you create the label first.

Switching to the web-based interface for Google Mail will take some work, but I've already found a compelling reason to make it worth my time. Using filters I can apply labels to messages as they come in. Then, I need only click the "Archive" button when I'm done reading to have those messages automatically filed. This is a terrific way to streamline the organization of my messages. When using Thunderbird, I'd have to manually move each message (or selected group of messages) to the appropriate folder. I would often miss the intended destination folder, and end up filing messages in the wrong location. Using Google Mail, I can ensure that mail ends up where it needs to go before it even comes in to me! All my mailing list messages are now labeled automatically, as well as mail from my family. I only ever need to click the "Archive" button now, which makes for a pretty easy to remember workflow.

I installed the FireGPG plugin in order to be able to use GPG signing and encryption from within my web browser (although I hardly ever encrypt anything, much to my disappointment). I'm now skimming the list of Greasemonkey scripts for GMail and will probably also investigate a few Firefox add-ons to see if there are additional ways to make my Google experience even better.

So far, I'm pleased with what Google offers, and I'm happy to get out of the mail administration game. I'll soon be able to downgrade the VPS service I use, since I no longer need the additional memory or disk for email processing and storage. I'm looking forward to sharing a calendar with Carina. And while I doubt I'll make use of the Google Docs service, it's nifty to know that it's available to me should I find a need.

DIY

For many years I've been an ardent supporter of the "do it yourself" approach for server management. I used to run all of my websites and mail servers from my personal DSL line using a computer in my spare bedroom. Earlier this year, I finally got fed up tending to the hardware and purchases a VPS account at TekTonic. This has served me better than I expected, and I've felt extremely liberated not having to fuss with hardware.

I'm growing increasingly tired, now, of managing my mail services. It's no longer interesting to me to keep up with the latest anti-span tactics, or to restart the mail services when they fail for whatever reason. As such, I'm seriously considering letting someone else manage my email services for me.

I know several people who use -- and rave about -- webmail.us. I'm comfortable paying for someone else' expertise in email management, so I'm willing to fork over a couple bucks a month for a reliable email system I don't need to babysit.

I also see that Google for Domains is available to me, and for free. I'm not entirely keen on the idea of letting Google hold all of my email (plus all my family members' emails), but I'm not so opposed to that idea as to rule out Google altogether.

A quick skim of some of the reviews of Google For Domains suggests that it should be satisfactory for my needs, and will offer a number of additional benefits which may or may not be useful to the family (shared calendar, for example). I think I might give it a shot...

Google Reader Trends

I just noticed that Google Reader has a "trends" link. I don't recall seeing that before; though I admit that I don't often look at the interface for Google Reader, as I'm too focused on the news content it is displaying to me.

From your 83 subscriptions, over the last 30 days you read 4,730 items

I haven't starred, shared, or emailed any items. I'm not much for social networking functionality, which explains why I haven't shared or emailed anything.

What's really interesting in the trends is both my personal reading habits, and the posting habits of the sites I read. I read more items between 7 and 9 AM than at any other time through the day. This makes sense, since all the news that has accumulated overnight is waiting for me. And for some reason, I read more items on Wednesdays than any other day of the week.

BoingBoing and Slashdot both post about 20 items per day. The feed for my Flickr contacts is running around 15 items per day, while the Flickr feed for the tag "skippy" is about five items per day.

Ultimately, all of these stats are completely meaningless, and not particularly useful to me. I don't really care how often sites publish, or what percentage of their posts I read, or when I read news the most. I care about reading the news, not reading metadata plotting trends about me reading the news. That said, I'm now interested in subscribing to a variety of feeds to which I wouldn't normally subscribe, just to see what sort of trends might develop.