Scribbling

I’ve been scribbling online for thirty years, more or less:

July, 1995 - vi

In 1995, Wolfram Research had a server dedicated to employee web pages.

Here’s an old log entry:

Tue Jul 11 17:04:46 CDT 1995.
Woowoo: I have a web page now.
I happened to be in David's cubicle, and he said, "Do you have a web page?"
"No," I said.
Now I do. It was easier than I thought: just create
~/public_html/index.html
and put html stuff in it. David did all the work, which makes it even easier.

I kept at it, over the next few years, adding pages for various topics: books I’d read, movies I’d seen; a blog-like section, titled Chronology; and a page of random thoughts, grandiosely titled Meditations.

These were all static text files. They were manually created & edited, directly in ~public_html, using unix shell commands and vi.

May, 1999 - A brief hiatus

In May of 1999, Wolfram Research reorganized its public web servers. The employee home pages did not survive the transition: at 5pm on May 14th, they all went offline.

I don’t blame the company for doing that. It’s not a hosting service, it doesn’t owe me (or anyone else) a platform.

The HR folks offered an Internet Agreement Form, where employees could request public visibility for their home pages. If your content were somehow relevant to the company’s business interests and you promised not to post anything that would get them sued, your ~/public_html directory would be added to the new Wolfram Employee Home Pages server.

I declined to sign the Internet Agreement Form. Instead, I deleted all my files. (I made a backup first. Thanks for asking!)

As recently as 2019, the server was still online. A few of the WRIfolk had pages there, but it’s impossible to say how many: there was no directory, no index, no search function. Google had a few search results, if you knew what to look for.

December, 1999 - FrontPage

In January, 1996, Microsoft acquired Vermeer Technologies, and with it Vermeer’s one and only product, a web site creation / management tool called FrontPage.

Microsoft filed off the serial numbers, did a quick rebranding, and released Microsoft FrontPage 1.1 a few months later.

I bought it, because it was my habit, through most of the late 1980s & early 1990s, to buy anything that Microsoft released. I bought all the upgrades, too: FrontPage 97, FrontPage 98, FrontPage 2000. I spent a lot of money on FrontPage, but I never used it for anything until December, 1999.

My old web site had been offline for six months; I missed it, and wanted to get back online. I was tired of using vi to edit web pages, and decided to give FrontPage a try.

The new site was hosted on Netcom; I edited pages offline, then published the entire site via FTP. (Here in the paranoid 21st century, just the thought of sending login credentials unencrypted across the internet makes me twitchy. It was an innocent time.)

It took a long time to squeeze an entire web site through a 56k dialup internet connection, but that’s all we had: AT&T didn’t roll out DSL to our part of town until 2004.

FrontPage had a notion of page templates, so I could have standard headers & footers; but I never could get it to work reliably. Often enough, FrontPage would quietly throw away all my meticulously-crafted templates and force me to recreate them.

I never did like FrontPage very much. (Perhaps that marks the end of my buy every Microsoft release phase.)

May, 2002 - CityDesk

In April of 2002, I started experimenting with CityDesk, from Fog Creek Software. It was a Windows app, with a nice page editor; and it was scriptable. I could write code to generate index pages & things.

It was a big win over FrontPage, though it followed the same edit offline, then publish paradigm.

It was also expensive. Fog Creek offered three price points: Trial, free but limited to fifty pages; Home, $99 but limited to five hundred pages; and Professional, unlimited but $350.

I started out with Home, that being all I could afford, and struggled to keep my site under the page limit - which was really a file limit: every html & css file, every image, every thumbnail counted as a “page”.

Fortunately, when CityDesk 2.0 shipped in August of 2002, Fog Creek dropped the Home edition. Everyone got Professional, with no limits.

November, 2002 - Pair Networks

In the early days of the internet, registering a domain was cheap: $10/year. I was oblivious, and missed the chance to register something simple & lucrative, like milk.com.

Later, registering a domain was very expensive: $120/year, or some such. I might have enjoyed my own personal domain, but not at that price.

By the early 2000s, registering a domain was cheap again, cheap enough that I was willing to pay for one. I should have taken my time, put some thought into it, and chosen something pithy or clever; instead, I panicked & went with patrick-rice.net.

(It’s a pain, having to spell that out to store clerks & other nosy people who want my email address. But I’m stuck with it.)

I registered my domain with PairNIC, and signed up for web hosting with Pair Networks. Aside from the personal domain, Pair offered more considerably more disk space than Netcom.

I made a few adjustments to my CityDesk site, hit Publish, and - two hours later - was up & running in my new home.

February, 2005 - MovableType (1)

Fog Creek promised that CityDesk 3.0 would ship by the end of 2003; it didn’t.

In 2004, Fog Creek assured users that they hadn’t forgotten CityDesk, that 3.0 would ship, Real Soon Now; it didn’t.

Eventually, Fog Creek stopped talking about CityDesk. They ignored users who asked about it, and deleted the disgruntled Where’s CityDesk 3.0? posts that appeared in the discussion forums.

By February, 2005, it was clear that CityDesk had no future. I poked around a bit, paid attention to what other people were using, and decided on Movable Type, version 3.1.

MovableType was written in Perl, a language I had once used but long since abandoned as vile & unreadable; but it ran on the server, and had a nice dashboard for writing and publishing from any web browser.

There were issues. The default site styling was ugly. (I don't remember what it looked like, but I do remember the complaints from my mother. She hated it.) It still followed the publish to static html paradigm that FrontPage and CityDesk had used.

But MovableType offered a way out of CityDesk, and that was worth a little difficulty.

November, 2006 - TypePad

Maintaining my own private installations of Movable Type (for the blog) and Gallery (for the pictures) turned out to be work.

The whole point of having a web site was writing, not fooling with other people’s software; so I signed up for accounts on TypePad and Flickr, and set about migrating all my data.

That turned out to be work, too. But I managed.

I’ve long since forgotten everything about TypePad: how it worked, what it was like to use it, what the published site looked like. It’s all gone. Perhaps I should have taken a few screen shots, for posterity.

Oops.

July, 2007 - Twitter

Twitter was a year old by the time I got on board, but my favorite username was still available.

I had fun on Twitter, though by the usual metrics my account was a failure: I didn’t follow very many people, I didn’t have very many followers. I had no clout, I influenced no one.

I worried that squeezing my thoughts into 140 characters would ruin whatever skill I may have had at long-form writing, but wrote 17,000+ posts … er, tweets … over the next sixteen years.

Then Elon Musk bought Twitter, and it wasn’t fun any more.

May, 2008 - MovableType (2)

By 2008, I’d been writing html for thirteen years. I was comfortable with it, and adding a bit of html markup and formatting as I wrote wasn’t a distraction.

It felt a bit like the old WordStar markup codes, back in the 1980s.

That was definitely a minority opinion. Plenty of people didn’t know html, didn’t want to know html, but still wanted to create web pages. (I don’t blame them. It's a question of focus: time spent fooling with html is time not spent writing.)

In May of 2008, the TypePad people rolled out a new post editor: no more markup, all WYSIWYG all the time.

I hated it. I wanted to write my own html, not rely on somebody else to do it for me. (<em> and <i> are not the same, damn it!) I had opinions about how to lay out the marked-up text. The new TypePad editor didn’t care what I wanted. It also had a tendency to lose posts, which is an unforgivable sin in any writing tool.

The TypePad customer service people assured me that the new post editor was very nice, and if I just gave it a try I might learn to love it as they do.

I was unpersuaded, and abandoned TypePad. (As a parting gift, the TypePad developers broke the export function, and held my data hostage for a number of weeks. Eventually, they fixed their problem & I got my data.)

I switched back to Movable Type, only this time I let Pair install & maintain it for me.

August, 2009 - Wordpress (1)

The second incarnation of Movable Type lasted only a year; and then I began to scheme for its replacement.

I settled on WordPress, even though it was written in PHP - another language I’d used in the past, then abandoned as vile & unreadable. (It may be that I am too opinionated about programming languages.)

Along the way, I switched hosting services, from Pair to Pair Lite. It was cheaper.

WordPress was good enough. I fooled around with themes and sidebar widgets; and I wrote thousands of posts over the next seven years.

In 2016, I lost interest. The posts grew shorter & less frequent, then stopped altogether.

Sometime after my final post (July, 2016), the WordPress site export function stopped working: it trundled a bit, downloaded a few MB of “WordPress Extended RSS” data, then died. I couldn’t get a complete backup.

The likely explanation is that it was running into some kind of CPU watchdog on the Pair Lite service.

I kept the site online for a few years. I updated WordPress whenever it asked me to; I even upgraded PHP itself once when WordPress complained that it was too old. But I never posted anything.

I pulled the plug on January 5, 2020 and deleted everything: WordPress, its database, and all its files. All that remains of my old web site is a single HTML file, hand-crafted just like 1995:

Nothing to see here. Move along.

I do have a complete backup, obtained before the watchdog started biting. Perhaps someday I’ll get all that imported here.

October, 2012 - App.net

People have been looking for alternatives to Twitter almost since Twitter began: a mass of dissatisfied users, careening from one site to the next, hoping against reason that this one would be different.

Most of Twitter’s challengers imploded after a month, or a year, and their users returned to Twitter. (One of them, Peach, went from hot new thing to pathetic has-been over a single weekend. To this day, their record stands.)

One that almost made it was App.net.

App.net was more or less the same as Twitter, except posts were larger (255 characters vs. 140) and it wasn’t free: a one-year subscription cost $50.

I desperately wanted in, so I could hang out with the cool kids, but didn’t have $50 to spare.

In October, 2012 the App.net subscription price came down (to $36/year), my disposable income went up, and I joined. (My favorite user name was still available, too. Always nice when that happens.)

I enjoyed App.net for a few months. It was a great place to hang out.

Alas, in early 2014 the business side of App.net fell apart: too many employees, too little revenue, no idea how to keep the servers running & the company solvent. The founders sacked everyone (including themselves), and started looking for a buyer.

The users all went reluctantly back to Twitter.

Before the App.net servers went dark - permanently - in 2017, I took a backup of all my data. From time to time I think about resurrecting the old posts, but so far they remain trapped in JSON like flies in amber.

March, 2014 - Ghost (1)

In early 2014, I was fooling with a new Raspberry Pi - perhaps it was the original YardCam project - and wanted to record my adventures. I had the notion to spin up a Digital Ocean droplet and install Ghost on it.

I gave it an old domain I had lying around - registered in 2010, mostly for laughs, but never used - and was off & running.

I wrote a few posts, over the next few months, then lost interest. At some point, I pulled the plug: deleted the droplet, returned the domain to its former limbo. (Yes, I backed up my data first. Thanks for asking!)

But Ghost was very nice - more appealing than WordPress, definitely - and I’ve been scheming ever since to get back into it.

May, 2018 - Wordpress (2)

Once again, I was keen to write, keen to get myself back online. I dusted off my joke domain, and handed it to Pair Networks.

The folks at Pair offered a new service: PairSIM, short for Software Installation Manager. Click a button, install & configure software on your server. I liked the idea of paying someone else to do all the installing & maintaining - the grunt work - and the price was right: $4.95/month.

(It wasn’t my server. I had a user account there, along with a few dozen other people. I wondered, sometimes, about all my virtual neighbors; but never enough to try to identify them or contact them.)

PairSIM offered a limited menu of applications: blogs, photo albums, wikis. I went with WordPress (for words) and PiWigo (for pictures), though neither was my first choice.

I had the notion to use Ulysses as my composition book. It synchronizes across all my devices, so I’d be able to write anywhere, edit anywhere, and then publish to Wordpress from anywhere. There were problems.

The Wordpress xml-rpc interface is notoriously vulnerable to brute-force attacks; the bad guys can throw passwords at it all day, until they find one that works, and Wordpress won’t do anything to stop them. Hosting services tend to rate-limit its use.

As it turned out, Pair Networks’ limit was somewhat lower than what Ulysses needed to publish to Wordpress. Even worse, the rate-limiter took the entire site offline for several minutes: no blog, no try-again-later error pages, no response at all.

Pair Networks tech support offered to exempt my IP address from xml-rpc rate-limiting, but I have too many devices - mobile devices - for that to work.

Forced to choose between Wordpress and Ulysses, I went with Ulysses.

I kept Wordpress and PiWigo up-to-date, but I never used either of them. A year later, I deleted them both & canceled my PairSIM subscription.

October, 2023 - BlueSky

Another Twitter alternative, BlueSky started getting some attention in … 2022? 2023?

BlueSky was in closed beta for a long time. I got on the waiting list, and … waited. Eventually, a generous fellow with a spare invite code took pity on me.

After the obligatory first post:

Hello, BlueSky …

… I set about looking for people to follow.

BlueSky reminds me of what Twitter was, before Elon Musk and the Nazis took over.

But after two years on BlueSky, time spent there feels … wasted? I’m not part of the conversation. My contributions go unread. What’s the point?

January, 2026 - Ghost (3)

...and here we are, in 2026.

I’ve been thinking - again - that I want to write, I want to put things online for people to read. I want to learn new things, and record the journey.

My demons have names like nobody’s reading and nobody cares. Perhaps I will ignore them, and see what happens.