I want to thank everyone, especially Prime members and Prime AF members, for their patience over the last few days as we’ve worked to resolved a number of technical issues tied to moving TPM over to a new server set up. Your memberships and your support are the bedrock of TPM and our team has worked at a fever pace to resolve what are fairly complex technical issues (for tech folks, basically interweaving authentication and caching). As of last night we believe we’ve resolved the issues with ads showing up for people who shouldn’t be seeing them as well as restoring comments. (Yippee! Appreciate the Congrats.) The remaining issue is some subscribers seeing partial RSS feeds instead of the full post RSS. We believe we will have that resolved too as soon as today, assuming no new unforeseen issues.
Whenever there are speed bumps like this people understandably ask, why did these changes make sense if things were working fine and then the switch made them work worse? Understandable question. The first part of the answer is that they won’t work worse going forward. These are annoying but very short term and temporary glitches we do everything we can to avoid but are hard to prevent entirely when you’re making such a big switch.
The more general issue is that we are always trying to find ways to make the site run as efficiently as possible at the lowest cost. Another fundamental balance is making sure we have software that handles aspects of TPM that are non-standard while creating as little of that software ourselves as possible — since maintaining software builds is highly labor intensive and we have a tiny tech staff. Just as important, part of the way we keep TPM moving forward is avoiding whenever possible VC-backed services which create dependence at escalating fees. In other words, we work hard to have as few tech needles in TPM’s veins (bloodsuckers, if you want to be sharper about it) drawing out your subscriber fees as possible. In general we look for robust open source software projects that we can lightly customize for our own needs to give us a good balance of customization, reliability and low upkeep/low cost.
The complexity of the whole thing is that the tech and business environment is always in flux. Prices of services or tech commodities are going up or down in different cases and different software — either commercial or open source or software we build in house — is becoming available or ceasing to be available. So you can have the perfect set up figured out at one point and then a few years later it no longer makes sense. Having done this for many, many years I can remember working with various old-school pre-cloud hosting companies, then saving lots of money moving to the Amazon cloud and now we’re moving off Amazon because their are now services that are cheaper while also (by specializing in the publishing software we use) relieving us of a lot of system administration work. Similarly changed circumstances for how we manage the front page.
In general since we’re not a prefab site, there is an arc in which early on we need to build things ourselves because they don’t exist or don’t exist with just the features we need. Then over time other services, and in some cases open source frameworks, emerge that allow us to improve and get out from under the software maintenance burden without incurring big new embedded costs.
There are many reasons TPM has survived and thrived for more twenty years. But this is a critical and from the outside at least likely an under-appreciated one. To run an independent operation like this one you need to sweat a lot of business, publishing and tech details to keep ahead of the game.
In any case, that’s a bit of a backstory about why we do these things. We truly appreciate your patience and understanding. And while we cannot always avoid these transitional glitches trust me that we make these decisions after a lot of careful thought and planning and always aim to use your membership fees in the most effective and efficient way possible to keep the focus squarely on the journalism.