Callum Macdonald

It seems like these are the relevant sections:

No “phoning home” without user’s informed consent. This seemingly simple rule actually covers several different aspects:

If the plugin does require that data is loaded from an external site (such as blocklists) this should be made clear in the plugin’s admin screens or description. The point is that the user must be informed of what information is being sent where.

As I read these, WordPress does not meet the same requirement. I can’t find a privacy policy anywhere as part of a WordPress installation. Further, as I read this privacy policy, it makes only passing mention of WordPress phoning home for updates:

For instance, WordPress.org may reveal how many downloads a particular version got, or say which plugins are most popular based on checks from api.wordpress.org, a web service used by WordPress installations to check for new versions of WordPress and plugins.

So back to my original point (I do concede that I was mistaken on themes), WordPress phones home, collects personally identifiable information, and does not disclose that to users or provide any kind of opt-out, never mind opt-in.

I do understand that WP / Automattic feel like the update service “creates value” for users, and I accept that for the greatest majority of users it does. However, it’s 100% possible to provide exactly the same service without collecting personally identifiable data. The site url is only included because it helps WP to gather better statistics, it serves absolutely no purpose to the user.

I feel like we’re somewhat going round in circles.

Yes, WordPress phones home, it does so without permission, without an opt-out, and it includes the site url. Plugins are forbidden from doing the same. That’s fairly simple, and it’s fairly obviously a hypocrisy.

The plugin Ts&Cs doesn’t say “no capturing of personal information”, it says no phoning home. Which is precisely what WP does, although we’re told that WordPress doesn’t store our details. Nonetheless, it does include a uniquely identifiable string, so WP.org can count how many installations are out there.

To say “it’s a feature”, that’s all it is, doesn’t change anything. Some companies describe DRM as a feature, while others describe it as a cancer.

In my opinion, the pertinent point is that WordPress phones home, without permission or consent, and uniquely identifies each installation when it does so. Yet plugins are explicitly forbidden from doing the same.

Wow, thanks for pointing me to the web interface. It looks like couchbase was not actually configured, I had to tell it where to store data, create an admin password, etc. I’m guessing all the cache attempts timed out, which is probably why page load times went so high.

Now it’s working, I can view data being cached, but page generation time is still a little higher than with straight mysql. Performance does look slightly better using the memcached plugin which implements getMulti() in PHP as I understand it, so that could be moxi at work.

I also tried uninstalling couchbase and instead installed memcached, very similar performance, still slower than mysql. I think at this point I’ll write off this exercise as a useful learning experience, uninstall memcached, and look at alternate performance enhancements.

Thanks a lot for all the input BigBlueHat, couchbase certainly seems more accessible now.

Cheers – Callum.

@bigbluehat: Thanks for the detailed response, I really appreciate it.

My page load time went from 200ms to 5 seconds. From reading the blog post it seems to suggest that there could be a 200% improvement using the direct SDK instead of the moxi proxy. However, in this case I went from mysql to couchbase and saw a 25 fold increase in time spent generating a page.

I tried removing this object-cache.php and instead using the one from the memcached plugin which implements it’s own getMulti if I understand correctly. Under those conditions, page load times went to ~3.5s. So memcached-redux = ~5s, memcache = ~3.5s, and mysql is close to 200ms.

I don’t have any kind of code profiler so I’m not clear where the extra time comes from. I’ll maybe try replacing couchbase with memcached and see if performance improves, but I’m guessing there’s a bigger issue here somewhere.

Cheers – Callum.