Since I’ve occasionally groused about the hosting service I get with Dreamhost, I wanted to be sure to mention when things go right, instead of only when things go wrong.

For a few months, the server that my account was set up on, and which hosts all three domains under my control, was being tweaky, resulting in sporadic downtime. I’d been building up a small library of downtime reports in my support page with Dreamhost, when last month, things took a turn for the worse…and then, soon afterward, for the better. Of course, I didn’t really know about the “for the worse” part until I got the news of the “for the better” side of things.

The last time my site went down, when I submitted the support request notifying Dreamhost of the downtime, they replied relatively quickly, letting me know that the server I was on had become “unresponsive,” and they were working on getting this back up and running. About half an hour later, my sites came back up, and I didn’t think much more of it. Until the next day, when I got the following…

This is just a note to let you know that we’ve moved your account to a new server! We apologize for the lack of notice, but this was an emergency move as a perfomance and stability measure.

Apparently the server I’d been on had gone seriously downhill, and had to be replaced, necessitating moving my account to new hardware (and hopefully all the other occasional downtime I’d had was due to the developing hardware issues, and I won’t get those again). Most of the rest of the message was covering anything I might need to do to ensure that everything worked as it should, and as it turns out, I didn’t really need to do much of anything, as the transition was seamless. Then, at the end of the note…

…again, we really apologize for the abrupt nature of this move! To try and make up for it a bit, we’ve also set your account to now have unlimited disk and bandwidth, forever!

Ooooh. See those last few words? That’s nice to see. “Unlimited disk and bandwidth, forever!” No worries about storing large files (not that I tend to do that), no worries about surcharges on the (very) off chance that I get Slashdotted or Dugg. Just keep paying my yearly bill, and I’m set.

Dreamhost, I’ve had issues with you from time to time, but this? This is good. Thanks!

As might have been implied by my last post detailing an evening’s work tweaking templates and installing plugins, I’ve decided to stay with Movable Type for my weblog. There are a few reasons for this, but it boils down primarily to two things: familiarity and loyalty.

This isn’t at all a slight against WordPress (which I was actively poking at), Expression Engine, or any other weblogging system, for that matter. I’m actually quite impressed with WordPress, and if I were starting a project from the ground up, I’d definitely include it in the list of strong contenders to run the back end. For this site, though, I decided that it was better to stick with what I knew and spend some time tweaking things than to jump ship entirely.

Right now I have a little over three years worth of experience with Movable Type (I switched over to MT from a similar but far simpler package called NewsPro on Dec. 21, 2001). While I certainly wouldn’t rate myself terribly high in the pantheon of expert MT users out there, after this much time fiddling and tweaking, I don’t think I’m any slouch, either. While I’m sure I could learn the ins and outs of a new system easily enough, in this case I’d rather use and build upon the knowledge I have rather than starting over from scratch.

Besides, in the time I’ve been using MT, the software itself has worked quite well for me. My battles over the past weeks have been with the comment spammers and their abuse of the limited resources of my server, not MT. Moving to another system might have worked temporarily, but it would only be a matter of time (and likely not very much time, at that) before the attacks started hitting that system — and I’m still not convinced that a PHP solution is the best choice for my webserver. Better for me to make a few concessions (disabling comments after 30 days, for instance) than put my server through the effort of serving up an entirely dynamically-generated website.

There’s one more big reason why I wanted to stay with MT, though — and that’s Six Apart.

As I mentioned above, I started using MT back in its version 1.something days, back when there was no Six Apart, just Ben and Mena in their apartment. Back then, I was one of many people occasionally popping up on the Movable Type Support Forums, and as often as not, it would be either Ben or Mena personally answering the pleas for help when one stumbling block or another was found. It’s things like that that add a more personal touch to software — and one of the reasons I’m fond of shareware programs like NetNewsWire, ecto, or many other programs where the developers are still personally involved with their user base — there’s the feeling of a real, breathing person behind the software, rather than a faceless corporation.

Obviously, as Six Apart has grown, Ben and Mena aren’t always as personally involved with their user base as they used to be. However, in my experience, Six Apart has yet to lose that personal, “real person” feeling, and that’s in no small part due to the excellent people they’ve been hiring, many of whom have been loyal users of MT for longer than I have.

When I got Slashdotted after news of my departure from Microsoft broke across the ‘net, I was using Six Apart’s TypePad service. As it turns out, I had the unenviable position of being their first Slashdotting, and those next few days became something of an experience (for both myself and Six Apart, I believe) in how to handle such an event. I’d already spent much of the day waging a losing battle with my inbox as comments, TrackBack pings, and e-mail missives deluged me, when suddenly iChat popped up with a friendly hello from Mena herself. I was a bit taken aback — it’s not every day I get an IM from the President of a software company, after all — but again, it’s things like that that impress me. Rather than assigning my case to one of the tech support crew, she and I spent the next few minutes working out ways for me to tweak the code on my pages to ease the load on the TypePad servers.

A few weeks ago, I realized that due to my own absentmindedness, I’d accidentally paid for a year of TypePad that I wasn’t going to be using, as I’d moved back onto my own server. It was a little frustrating, but I had noone to blame but myself, and said as much when I grumbled about it here. Imagine my surprise, then, when I got an e-mail from Brad Choate, who’d come across my post, pointed it out to someone at Six Apart, and had made arrangements with Brenna to refund me that yearly fee. I hadn’t asked for this, and there was absolutely no reason for Six Apart to do this for me — but they decided that it would be a nice thing to do.

Then, just a few days ago, Anil Dash noticed that with my battles against the spammers I’d started looking at WordPress, and he sent me a friendly little note asking if there was anything they could do to help me with my MT installation. I let him know that my limitations weren’t with MT, but with my webserver (and was barely able to keep from mentioning how nice it would be to find an Xserve PowerMac Mac mini on my doorstep one day — it wouldn’t have been at all serious, but I don’t know if Anil stops by my page often enough to catch my sense of humor), and thanked him for his note. Again, this is the kind of thing that impresses me — sure, on the one hand, he’s “just another blogger”, but he’s also the Vice President of the Six Apart Professional Network.

What it boils down to is that over the years, time and time again, I’ve gotten incredibly friendly and personal service from the crew at Six Apart. I can’t think of a better way to build and maintain customer loyalty than that.

So, to Ben, Mena, Brad, Brenna, Anil, and all the rest of the crew at Six Apart — thanks, folks. Keep on rockin’. :)

Mike is doing some brainstorming on how to predict and cope with bandwidth spikes when a post or page suddenly becomes a popular destination.

When a blogger’s work becomes successful enough to, for a moment, graze the underbelly of commercial publishing, it threatens the very low-cost predicate of the publication itself.


Setting aside for the moment the absurdity of the situation, which is clear, it seems to me that over the past few years we’ve seen this exact phenomenon occur over and over again. I’m guessing, now that media people have integrated the blogosphere into their information gathering practices, we’ll see it with greater frequency and to more devastating effect over time.

My bandwidth as of 11/23/03As I recently discovered, this is a very real worry. I’d joked in the past about the “perfect post”, that one blog entry that suddenly exposes a site to the world and brings in all the traffic that so many people wish that they had — but actually stumbling upon that “perfect post” has made it very clear just how much of a double-edged sword that can really be.

In Mike’s ruminations on how things like this can be coped with, he mentioned something that sounded like a possibility…

…I think there is a proactive business opportunity for the right business to defray these transient bandwidth costs, probably in the form of short term ads on the sites that are experiencing the bolus. […] I will note that it might even be cooler yet if this feature enabled Google keyword ads. Maybe it should be an independent service, or a program that the keyword service provides for bloggers, who are currently more or less specifically discouraged from using it.

I applied for Google AdSense at one point, but they turned me down. While it was a bit of a bummer, it wasn’t much of a surprise, as Google doesn’t seem to want to accept most weblogs into their AdSense program. It seems that if you run a very tightly-focused weblog on a specific topic (such as PVR Blog or Daring Fireball) you’ve got a good chance of being accepted, but less-focused weblogs (such as mine, yours, the one you’re going to read next, or the other 99% of the blogosphere) will be denied. Unfortunately, the exact methodology or reasoning behind the approval/denial process is more than a little unclear.

There’s a far more serious problem with AdSense, though. The approval system is capricious, even arbitrary. It’s understandable that Google wants to make sure sites aren’t just ad farms, and it’s in everyone’s interest that quality be maintained, ideally by human verifiers. Nobody wants to see those sad Red Cross PSAs that take the place of house ads on poorly-indexed sites.


The human verification process at Google, though, is uncharacteristically opaque. I’d assume they factor in the ads which would run on a site before approving or denying an application, and if I take a look at , I see some of value. Ads specifically targeted to weblog software, Manhattan computer repair, New York hotels. These all seem relevant and valuable to me, but I’ve been repeatedly rejected.


It’s not just sour grapes on my part. Take NYC Eats, a great little niche weblog. Aaron’s brilliant little AdSense senser shows , which makes sense since the letters “NYC” by themselves cost two dollars a click. But no AdSense approval there. The problem is the wording in theprogram policies:


In general, we do not accept personal pages, chat sites, or blogs into the AdSense program. However, if a site contains targeted, text-based content and/or provides a product or service, we may consider it for participation.

In a perfect world (well, my perfect world, that is), of course, Google would open up their AdSense program to the weblogging world at large. While their AdSense ads might be a little random on the main page of a site due to the random nature of the main page posts not giving clear, concise keywords to work with, if a site design includes individual archive pages than each individual post should have enough keywords to target a specific ad category (my Mac-specific posts would get Mac-centric ads, my political posts would get political-centric ads, and so on).

If they don’t want to do that, though, what if Google set up an agreement with TypePad (or other for-pay hosting sites) in which, in order to offset the cost of bandwidth spikes, Google AdSense ads could be (semi-)automatically added to a site when they reached a certain bandwidth point (90% of their available monthly bandwidth per their agreement, for example)? Each auto-generated template could include code something along the lines of <$MTAdSense><!-- include "/ads/google/" --><$/MTAdSense$> that would be automatically triggered by the TypePad servers when bandwidth exceeded whatever the cutoff point was. Any revenue generated by clicks on the ads would automatically be siphoned to TypePad and applied to offset the costs of the extra bandwidth usage during the spike.

There could even be a toggle in the TypePad preferences that allowed a site author to insert a “registration key” if they were accepted by the Google AdSense program that would enable the AdSense ads on a full-time basis. In this case, Google would send any revenue to the site author as per their usual setup, instead of sending it to TypePad.

Just an idea. Workable? I haven’t got a clue — barriers include the coding of the feature (while I’m no program-level coder, it doesn’t strike me as being too terribly difficult of a feature to enable), inclusion of the feature into already-existing weblogs (not difficult for TypePad Basic, Plus, or Pro levels using the auto-generated templates, Pro levels using advanced templates would need to add the requisite code themselves), and — most importantly (and possibly most difficult) — Google and TypePad (or, of course, whatever other hosting service that might be interested) negotiating the partnership. Still, if it could be worked out, I think it could be useful and beneficial to the blogging community at large.

Since getting Slashdotted, my bandwidth has been going through the roof. On an account rated for 5Gb/month of data transfer, that on a normal month would use roughly half that amount, I hit about 90Gb of data in the last five days of October, and over the first four days of November have already hit about 40Gb of traffic. Crazy.

After realizing this, I did a bit of investigating, and realized that each of the posts that have been getting the most traffic (Even Microsoft wants G5s and Of Blogging and Unemployment) have picked up so many comments that they were up to ~300kb each! At that rate, each page would only need to be loaded three times to produce 1Mb of data transfer — and with the amount of traffic I’ve been getting, that number grows quickly.

In an attempt to try to slow things down a bit, then, I’ve had to both disable any new comments on those posts, and disable the display of the comments I’ve already received, which brought each of the two pages down to around 60k. Turning TrackBack pings off brought the page size down even more, to around 6k each — far better. If things die down, I hope to be able to re-enable at least the display of the TrackBack pings, if not the comments (some of them are pretty entertaining, if not rational) next month sometime. Until then, though (and quite possibly permanently), they’ll have to stay ping- and comment-free.