Friday, August 09, 2002

How we're spending our time at Pyra

How we're spending our time at Pyra

The original charter of this blog, as explained over there on the right, was to tell not only about known issues, outages, etc., but also what we're working on with Blogger. That's an important thing that I believe in doing�and would appreciate at a customer/user�but in the day-to-day stress, gets knocked down on the priority list. So, here's an update:

Sometime later this month, the first public, large-scale, non-Pyra-run installation of Blogger will go live. It's a unique partnership/licensing deal that we've been working on for a few months. It won't have a direct effect on the Blogger usage of most people reading this. However, it's important for a few reasons: To implement this deal, we have finally finished the long-in-the-works upgrade to an entirely new codebase (no more ASP). You see, a huge bottleneck for the last year-and-a-half when it comes to fixing bugs and adding features is that much of the code we have in production currently we knew we needed to move away from in order to go where we wanted to go. So there have been a lot of quick fixes or band-aid approaches or don't-know-how-solve problems because we can't invest in the code and get new stuff written at the same time. That, and much of the current code was just sucky and when we started writing it, we of course didn't have nearly as clear a picture of how it should work as we now do.

So, we've got an all-new and improved base on which we want to actually invest (and that was written by people currently working for the company). That's huge. And it's being thoroughly tested right now (as I write this, in fact) by the QA (quality assurance) department of a big company who is not going to be easy on us about things breaking.

The next task, of course, is to move this code into production on Parts of it already are. The new publishing engine in Pro, for instance, is related to that. There are currently some issues with it, which we're figuring out, but it's in general much faster and better. And, importantly, it clears the way for some bitchin' new features and improvements. But the main transition of the interface (both free and Pro) and everything will start happening within a couple weeks. This won't be painless. Replacing a system that tens of thousands of people, with all kinds of different client and server-side configurations cannot be painless. But we'll try to make it as smooth as possible (and since things aren't exactly error-free right now, it'll probably be an improvement no matter what). Also, the database, behind everything, will stay the same, so we can have concurrent versions running at the same time.

In the infrastructure realm, we're continuing to make upgrades that perhaps aren't very noticeable, but are, nonetheless, a big improvement. Our system administrator (yes, we have a dedicated system administrator now�let's here it for Rudy!) has been kicking some ass in making things more reliable and faster. Reliability and performance, of course, have never been our strong point. And Rudy didn't have an easy task coming in, but things have been getting better. And considering that we're handling an every increasing load (about 1,800 new user sign ups a day right), and he's had virtually no budget, that's pretty good.

One of the things Rudy did recently, for instance, was move Blog*Spot to some new servers, which are now redundant and handling the load much more better. Though it's not as complicated as Blogger, Blog*Spot is an increasingly large percentage of our traffic, and the one server it was on was definitely at its limit (or beyond on some days). Not only is it more reliable now, the new setup allows us to add more servers as needed.

What else? Well, speaking of infrastructure, something that plays a major role in determining all of this, of course, is the company infrastructure. Besides Rudy and myself, there are currently four other people spending most or all of their work time on Pyra stuff. With the customer service person we're about to hire, it'll be five. And there are a couple other developers working on projects for us on the side.

Except for the couple other developers, who don't get paid at all (yet), the thing that's paying all these people, more or less (in some cases, way less ;), is the revenue from Blogger Pro, Ad-Free Blog*Spot and pyRads (in that order�by a long ways). Plus the licensing deal I mentioned above, which will be a significant, but not huge part of our income this year. And that's the way we like it. Jason and I have determined that we definitely want to focus on tools and services developed for and paid for by end users. It's the most fun, least frustrating, most scalable way to do cool stuff when you're company as tiny like ours (that doesn't always want to be quite so tiny).

Anyway, that's a rant for another time. The point is, we're making progress. Sometimes I think we're not, but we are. Sometimes I'm banging my head against the wall wondering how we're ever going to get done the things we need to (let alone want to) get done. But then I remember that at the beginning of this year, I was the only one in the whole company, and we had virtually no revenue. And I wasn't having much fun. And then I think, yeah, things always take longer than you think. But we're making progress.

So our plan, if you were wondering our plan, is to improve our current stuff and build more useful services and tools that you will like and some of which you'll pay for. We hope. We're pretty confident, actually, that enough people will.

And thanks for your support. And I don't mean that in a Bartles & Jaymes sort of way. Because those guys weren't real. But I am thirsty...