I started writing this as notes for my own use, and wasn’t really planning to post it publicly. However, I didn’t find any comprehensive google exit plans that were suitable for people in my position, and it seemed like an interesting area for discussion, so up it goes.
While making my regular Google backups (detailed below in “Backup ALL the Things”) over the weekend, I decided it was time to update my plans for bailing out of google’s services if necessary, and discovered that there may be superior alternatives to some of the services I’ve been depending on. Google’s vast infrastructure, development resources, ubiquity and integration have tended to make them better than self hosted options. The fact that they are a single party who has thus far been generally responsible with user data makes them more attractive than other hosted solutions. Both of those situations are subject to change.
- Google is getting increasingly intrusive, and I’m not sure if I’m OK with that.
- Plus is fragmenting my on-line presence, and not giving me much in return. Also, plus kind of sucks for the things I want to use it for – limited markup and one link per post makes for a terrible discussion platform.
- I prefer to self-host and be in control of my data.
- I’d rather not risk losing services and data I depend on over a ToS dispute (see Nymwars).
- Lots of other versions of explanations as to why one might just want off the google parade.
- Backup ALL the things!
This should be done regularly anyway, just to protect against account lockout, provider data loss, or other sudden problems. I try to do monthly, but am usually behind.
- Pull an IMAP backup of GMail. I do mine with getmail configured as suggested here, which is great except for requiring a plaintext password in its config.
- Pull a takeout copy of your data from whichever google services you use. (Good on Google for providing this – it’s one of the primary reasons I use some parts of google’s ecosystem).
- Pull a copy of your reader subscriptions with the built-in export mechanism.
- Find suitable replacements for used services and software.
There isn’t much point in going to alternatives that aren’t self hosted, it just shuffles the problem and further spreads your data around.
Replace reader with a self hosted synchronized feed aggregator like TinyTinyRSS, which seems to be the most promising, or rssLounge which also looks pretty good. They both seem to be a little less featured than reader, but reader is becoming steadily slower and less functional, so the loss is decreasing.
This doesn’t solve the problem of wanting a system for sharing – presumably to the blog. Since Plus is… actually pretty terrible for the things I want it to do… it shouldn’t be that hard to do better. I think it should be pretty easy to build a RSS reader -> blog work flow using one of WordPress’ push publishing mechanisms.
For pushing my content, I already have a hosted blog where I put original content (ie. this, here), and anyone using Plus probably uses an RSS reader anyway. The same applies for most people I am interested in hearing from.
Admittedly WordPress isn’t the greatest blog/CMS thing, but it’s well supported and easy to manage.
- Photo Upload:
I really, really like automagically pushing images from my phone to the ‘net. Flickr, Dropbox and a number of other services have apps to do it, but I can’t find one that is “Upload to my computer” instead of “Upload to someone else’s computer.” Worst case, it should be straightforward to do it with a script using rsync over ssh or a push-pull arrangement with an external service.
Like most people, I have half a dozen email accounts. Right now it all forwards to gmail, but email is email, I just need to swizzle the forwarding paths to change which server actually holds on to all of it – I would probably go to my pappp.net account since it is the one most in my control. Picking Horde, Roundcube, or Squirrelmail for webmail is literally a click away, and there are plenty of IMAP clients for use on my own machines.
Sadly, none of the alternatives are quite as good on the features front: The substitutes for full-content search mechanism and n-to-m labels (at least there are finally other implementations of that system) are inferior at this time, at least as far as I know.
There are some things google is better at, but duckduckgo, mooter, Wolfram Alpha and Bing all have domains they excel in. At the very least stop searching logged in.
Google has DNS servers at 18.104.22.168 and 22.214.171.124. They are fast and easy to remember. OpenDNS has a pair at 126.96.36.199 and 188.8.131.52 that are not so easy to remember, but just as fast and don’t have the same kind of privacy concerns. There are plenty of other sources – DNS is DNS, it is just a matter of speed and convenience. I currently set search order mostly at random, but it wouldn’t be hard to demote google’s.
- Chromium Sync features
I don’t know of any self-hosted option with similar functionality, but I’m not that dependent on it anyway. Worst case I can just pass bookmarks files around with scp like the [good|bad] old days.
It’s all XMPP, and google supports federation with Talk, so it will inter-operate with any other XMPP network. I’m not sure which other network I would use, but I’m much less dependent on chat than I used to be, and very pleased that it has mostly come around to an inter-operable standard.
- There is also the client software situation:
- Decide if Chromium is an issue – Chromium is just a sleek WebKit based browser once logged out of the google integration, but if that ceased to be the case it wouldn’t be that bad to go back to Firefox.
- Android is the best of the worst, and the alternatives for a reasonably open mobile platform are dwindling fast. Maemo/Meego/Mer and WebOS are dying breeds with limited hardware options, and the OpenMoko stuff is eternally useless. iOS and Microsoft’s current mobile incarnation are totally unappealing in a wide variety of ways. I think Mer on the most suitable bit of hardware it supports is probably the best bet, but I’m pretty sure that home-brewed Android ROMs (Cyanogen & co. or even replicant if things get ugly) will stay as OK as a smartphone can be for some time to come. I like the connectivity and convenience of a smartphone, but I did dumbphone+Wifi connected handheld for a long time and could go back if necessary.
- Find a way to import the liberated data
- I don’t know of a canned way to integrate old reader shares and Buzz/Plus posts into wordpress (or whatever CMS). The HTML export looks doable with some scripts, but would be complicated because of crossposting and format integration. Pulling in the comments would be a bitch, but amazing – the data is there just hard to match up.
- I think I can imapsync the unarchived gmail mail into my hosted account, and the archived stuff can rot in an archive.
- Moving contacts is easy – the takeout operation exports vCard format contact information which is an open standard and easily imported into almost any other email system.
The short answer is “not quite yet.” There are basically two things that would cause me to bail out of Google’s services – better alternatives, or bad behavior by google that outweighs the value of their services.
The latter hasn’t happened yet, but is creeping ever closer as they become more and more intrusive.
The former isn’t even close on some fronts (ex: Gmail’s feature set is enormously better than anything else I’ve used), and is rapidly approaching on others (Reader’s increasingly poor performance as a news discussion platform).
So: There is my plan for bailing out of google, mitigating the risks inherent to depending on google, and a list of google product alternatives likely to meet the needs of users in similar positions. I’m leaving out a few minor things (I occasionally use google as an OpenID provider, and Hangouts are the best video conferencing mechanism I’ve ever seen), but think I could be off google with a minimum of disruption in a matter of days if the need arose. Comments, suggestions and similar plans you would like to share are more than welcome.