Publishing to Smugmug, a geek’s view (Part 5)

Welcome to the final piece in this series on geeking lightroom and Smugmug.

To close this series out, I wanted to talk a bit about Smugmug smart galleries and ways I’d like to see them improved.

In practice, the implementation of Smart Galleries is pretty good. Because you can chain smart galleries together and create a gallery with interim results to use to create a final gallery, there’s very little you will be unable to do if you’re willing to think through building your selections and exclusions.

I ran into one bug which is fairly easy to work around: that a smart gallery with multiple EXCLUDE selectors doesn’t seem to work right. it almost looks to me like the 2nd EXCLUDE causes the system to throw all of the exclude clauses out. Other than that, the system worked as I expected it to.

It would be nice to see the size limits (1000 images for a smart gallery, 5000 for real galleries) increased, but to be honest, I can’t honestly say I think it’s a significant limitation. In practice, I only ran into this first limit twice, both for galleries building interim selections, and in both cases, it was easy to find an alternative way to solve the problem.  3,000 and 10,000 would work for me for years. 1,000 and 5,000 work fine for me as long as I remember to think about them.

I would like to see more than five selectors on a smart gallery (ten?). It would reduce the need to create interim result galleries. I would also like to see smart gallery selectors get smarter and allow a more flexible logic, especially mixing AND and OR logic (“include (fred AND jane) or (jerry and elvis)”). There’s a tradeoff here: allowing more complex logic is going to create more opportunities for people to get themselves into situations they don’t understand why it’s not working.

I would like to see some new ways to do selections, such as “all images in any gallery in the ‘birds’ section”. Alternatively, I would love to be able to create a smart gallery of all images in “this gallery” that are not shown in any smart gallery (the “which images are missing” gallery). One place I had to think through doing things was my mammals: there’s no easy way to say “all mammals, except “these” and “these” and these”) because of the 2-exclusion clause bug, and then soon after, the 5 rule limit.  ended up solving it by re-arranging my keyword nesting in Lightroom, but that means my sea otters and sea lions aren’t listed as mammals. It’s a minor thing, but it annoys me a bit, because I’m kinda picky at times.

This all starts looking like a relational database schema after a while, and I have to sometimes resist the urge to over-normalize the design. But if you think about it, smart galleries look a lot like views, and the selection criteria start looking like triggers.

I’ve noted a couple of things that people ought to be aware of before they start doing this.

First, all of your stats will show up as being tied to the real gallery, so you won’t be able to see analytics on which smart gallery the image was actually seen in. If you want that kind of data, you’re going to be unhappy, because there’s no way to get it. For me, this is fine. If that kind of detail on how people are finding images is important to you, this strategy won’t work.

Second, I have noticed a few cases where updating the galleries locked up or broke the site for short periods of time. It looks like while a smart gallery is being updated by the hidden background processes there are some locks put in place or the data is inconsistent. These inconsistencies tend to be very short lived, but they are visible to the end user. This implies that if you’re doing significant updates that affect many of your galleries, you might want to split them up into smaller chunks, or time them for when things are quiet (like start the publishing push before going to bed). If absolute 100% “my site is never funky” is a requirement, this strategy might not work for you. In all cases where I did see this, it seemed to involve significant sets of updates (> 100 images being pushed that rippled across the galleries) where it seemed like the gallery update daemon was trying to update the galleries while a previous run was still updating them. And in all cases, the glitches fixed themselves and went away within a few minutes of the publishing push finishing. And on the vast majority of publishing pushes (95% of them) I never saw a thing, so this is a minor case at best.

Overall, when I started this, I wasn’t sure if Smugmug’s smart galleries would hold up to what I was trying to do. I found they handled the abuse I threw at them just fine. Now that I have this up and running, it’s my greatest hope I won’t need to re-think how I publish things for at least 3-5 years; and rearranging the furniture using this strategy is easy and mostly painless. It takes a little bit of work to get set up; it saves a lot of work over time. I can live with that kind of investment.

You might also want to read:

1. Publishing to Smugmug, a geek’s view (Part 4) Welcome, as we start winding down this series on geeking lightroom and Smugmug. So, now that it’s all ready to use, what sort of things...
2. Publishing to Smugmug, a geek’s view (Part 3) We continue looking at how to set up your lightroom and Smugmug collections to work together in ways that will make your life easier in...
3. Publishing to Smugmug, a geeks view (Part 2)   In part one of this series, I walked through setting things up so that Lightroom and Smugmug were cooperating in getting images published from...
4. Publishing to Smugmug, a geek’s view (part 1)   I have been off under the hood of my publishing setup working out the gory details of how to get all of my former...
5. Why the geeks were wrong about .Mac…. WWDC Predictions – O’Reilly Mac DevCenter Blog: Our curmudgeonly java.net editor Chris Adamson, started things off by simply offering up a repeat of his last...