Where Some Ideas Are Stranger Than Others...
A Moonspeaker Retrospective
One day way back on 30 April 2003, the Moonspeaker made its first appearance on line. It reflected its times in its initial incarnation, which was a frame-based layout with a left hand navigation menu after an initial selection from the initial splash page. The initial splash page is illustrated below, and if you really want to try having a look around on the old version, the Wayback Machine will oblige via its archive of old alexa captures (or you can check out the Moonspeaker's website summary there, it's interesting to look at though again the documentation is poor). That's what I had to do myself, because it turns out that I didn't even have a crufty back up of the Moonspeaker any older than 2009. Today the appeal of frames is difficult to understand, although I suspect that to begin with they were about mimicking aspects of playpen web browsing environments like america online. The original site you started from would still be there, and you could navigate all over the web in the main frame. People soon grew irritated with this, including myself because it wreaked havoc on bookmarks and sometimes generated terrible confusion about what pages were hosted where. The original Moonspeaker constant graphic was not too different from the present one, just laid out on the vertical. Admittedly it took awhile, but I did finally conclude that "splash pages" are in fact "go away pages" even if they don't have light text on a dark background, but it did take awhile. At first everyone just accepted them, I think because again they seemed familiar because any program you started up on a computer put up a splash page with or without some minor animations to distract you while it loaded. Since the early web with images could take a noticeable amount of time to render, the parallel wasn't unreasonable to begin with.
The first mode is of course "web site sucks because the designer is a beginner and/or trapped in frontpage." It happens. The second is, "pretty good, isn't selling anything, and reflects an inclination toward utility over artsiness." The third is, "if you're selling something, you'd better not even think of putting light text on a dark background." The fourth is then, "hey, you're an artsy type – go right ahead and use light text on a dark background if you insist, especially if you aren't trying to sell anything. Even less guarantee than usual that folks will poke around though." This still strikes me as a decent broad strokes assessment, even in this age of overused wordpress templates and remarkable sharepoint survivals. Depending on how my experiments were going, I think it is fair to put the early Moonspeaker among those sites in modes one or four. Such is life.
On considering the lower right hand corner, there are two more items of at least minor historical interest. The "Made With Macintosh" graphic was once something of a big deal to have on your main page. Apple was in trouble still in those days, and it looked like there might be an effective monopoly in the consumer market of machines running windows very soon. Luckily that didn't happen, as that would certainly have made the present day problem of software vulnerabilities and privacy invasions that much worse. The second graphic reflects deep internet prehistory, before the days of spam filtering. Not only was it absolutely necessary to obfuscate your email address or watch it collapse under increasingly desperate offers of penis enhancement for only a small amount of money (pre-viagra days, even!). There was actually a good samaritan service being run that redirected spam bots into a data sink hole that filled them up with crud. The logic being, the lists of email addresses would become so corrupted by garbage they would finally bail out of the whole sordid business for something less obnoxious or sadly more effective. I suspect it never really worked, but it was worth a shot.
I was basically broke in this time period, so apart from options like netscape's composer, I had no access to any medium to full-service web authoring software. Netscape's composer produced some of the worst html possible until frontpage came along, to the point that even once I was familiar with html I could hardly make sense of it. So I ended up coding the original layouts and general content pages in the minimal text editor of the macos system 7 to 9 days, simpletext. Just as its name says, this was a minimalistic program. No colour, no text formatting, no line numbers. By its later incarnations some minimal text formatting was available, analogous to the surprisingly expensive minimal word processors available ostensibly to help people write on computers without being distracted. To begin with simpletext's main purpose was to allow a person to open the ubiquitous README files that explained how the early macs worked. Believe it or not, this really wasn't too bad as a web authoring tool. The Moonspeaker was pretty tiny in those days, so the as yet undreamed of "content management system" was unnecessary, and writing web pages by hand hadn't been relegated to the virtual equivalent of despised manual labour yet. This screen grab is also a good example of the different metaphors being used to guide the original appearance of application windows. The ridges at the top of the window are grab bars, meant to suggest it is possible to drag the window around by them. The lower right corner has some ridges too, for resizing the window by. Hilariously, it was ages before I realized that the box in the upper right corner could be clicked to create a split view. In any event, for the first two years or so of the Moonspeaker's existence, I might have anywhere from 5 to 25 simpletext documents open, trying out different things design-wise and flailing about still unsure just what the Moonspeaker was about.
Screen grab of the german version of SimpleText, courtesy of Wikimedia Commons.
The Moonspeaker didn't look much different, but I had made some changes by then in terms of tools for editing it. At the beginning of 2004 I had switched to the free version of bare bones software's editor, bbedit lite. It made for an important workflow change and facilitated considerable code clean up which was definitely needed after so many design and content experiments. I ended up carrying out a full site republish, and this seems to have drawn some attention from the alexa web crawler. According to the wayback machine, in 2003 the Moonspeaker was crawled only once, but in 2004 there were 8 captures. I did post a bit more by way of fiction and other writing at the time, but based on the nature of the site and how often I managed to update, a combination of both frequency of updating and number of links to other sites dictated how often a snapshot was taken. Since the Moonspeaker still didn't have a very strong identity, making plausible links to other sites was difficult, and I lacked the patience to render my rangy collection of bookmarks into a "cool places to visit" page. Accordingly, there were no snapshots for 2005. The original Amazons section was the main constant, providing access to a very early draft of a still developing writing project. In the meantime, I was busy having fun with a text editor that had code colouring and line numbers, and building the odd website for other people. These were generally ephemeral sites tied to specific events, such as conferences, and all done on a "will work for experience" basis since I did need more design practice and they were inclined not to pay much.
By 2006, I had added some more material but had just finished a post-secondary degree and started a real job, which meant the possibility of actually having an apartment and a computer that wasn't five years old, not necessarily in that order. I had a tiny blackboard on my wall with reminders and ideas on it for posting to the Moonspeaker, and did basically nothing with it. The splash page concept had gone moribund on the web, and it was evident that the Moonspeaker needed a significant overhaul and I needed to sit down and work out what the hell it was for once and for all. Luckily my new job happened to be: a web site design and production gig. I had no idea then, but this was a more important year from an information management perspective as applied to websites than I knew, because it was the year I stumbled on the first edition of Rosenfeld and Morville's Information Architecture for the World Wide Web (the link is to a page for the second edition). It's a cliché, I know, but this book did change my life. Besides completely altering my perspective on how to design the background structure of a website, it gave me the tools to handle the dawning age of near information glut and reintroduced me to the wonders of libraries. Firmspace libraries are among the first major information architecture and management challenges humans have created, and the principles are not so different between media. For the Moonspeaker, it got me thinking much harder about the relationship between site visitors and site navigation. I also learned how to count how many contractors had successively revamped a website by counting the different navigation methods provided. There can conceivably be a good reason for there to be two for a huge site, but when you run into a site with more than three, chances are you will have to use an external search engine to find anything.
2006 was another year of editor changes though, in part because bare bones software replaced bbedit lite with textwrangler. The reason for the name change remains unclear to me, since the relationship between free textwrangler and paid bbedit remained unchanged. If you wanted the additional features available in bbedit, you needed to cough up some cash. But in the meantime the free editor was pretty good. As the new apple logo shows, this was also an editor version that could not run under the old time macos. If memory serves (sometimes it doesn't), I was running macos tiger by then for sure. Apart from a big site republish, I had worked on linking the Moonspeaker up more with the rest of the internet, hence the new graphic from the news website rabble.ca. At the time rabble.ca had been through a period of exceptional notoriety due to false claims that not only was it a terrible left wing rag, it was funded by lots of federal money. It wasn't then and isn't now supported by government money, and it has drifted rather awkwardly between a sort of left wingness and a liberal individualismness from year to year since then. For the Moonspeaker, the spate of updates and new links meant it got captured by alexa 7 times, almost reaching its previous all time high. After this year though, my explanation for what makes a site eligible for crawling fails. The apparent correlation between number of updates and external links vanishes. I find this a little disappointing, especially since there does not seem to be any explanation of what triggers alexa to crawl a site at all, perhaps in an effort to prevent any return of "search engine optimization" nonsense. (I don't want to try to "game" the alexa crawler – way too much work, batman – but it would be interesting to better understand the criteria and think about whose priorities they reflect.)
Screen grab of an early version of bare bones software's textwrangler. Courtesy of the lowendmac.com
Please bear in mind that this is not a reflection of the behaviour or documentation of how the wayback machine's criteria for archiving sites. In that case, once the captures are by the wayback machine instead, its FAQ provides some helpful information. The FAQ is quite clear first of all that only publicly available sites are archived, which makes sense. A robot hardly has time to deal with passwords and so on in the particular context of a not-for profit archiving project. Another key point is that more crawls are done of a given site if it has been linked to by other sites, so in this case it is clear that the more other sites link to the Moonspeaker, the more often it is likely to be crawled. The FAQ also states that the number of crawls does not reflect the number of site updates. Based just on the 10 crawls carried out in 2016 on the Moonspeaker, it appears that there are at least three other crawl types involved here: Web Wide, domain Survey, and HackerNews, all of which appear to be part of the internet archive's overall activities.
The later version of textwrangler introduced a more common interface approach these days, the document drawer. I kept using it right into 2009, when the Moonspeaker finally underwent a serious and long overdue major redesign, illustrated below. There's a lot going on that follows up at last on things like actually indicating when the last update was and the like. The late version of textwrangler facilitated many changes such as auto-updating using scripting and insertion of code snippets. Prior to the completion of the 2009 redesign, one thing I had begun to appreciate was that the splash page actually hid more than it revealed. This reflected more than the section labels, which were only so helpful, although this is still a common failing of websites today and still something meant to encourage people to look around. There's simply no simple way to feature or even reveal new postings without redoing the entire layout each time. I had also painted myself into a template corner by late 2008, because I effectively had a different template for each major section. That way madness lies, even with the help of an expensive content management system. It's bad for making new pages and worse for visitors who can never be quite sure where the site navigation is going to be or how the information is going to appear on the screen. The importance of not having the site look completely different from moment to moment became much clearer to me in 2008 because that was also the first year that I stumbled over an A-B test being run by amazon.com. One day the clunky but familiar interface was there, the next day it wasn't, and the day after that it was back again. Since this was a site I shopped on, and web-based fraud was a growing risk, such changes made me a very nervous potential customer. Sure, the Moonspeaker doesn't sell anything, but that doesn't counter the point that a site that varies unexpectedly puts visitors off.
But now let's have a look at the 2009 redesign proper. It's a big one, and should look pretty familiar. At long last, the splash page is gone for good, first of all. There are some nice easy update parts, and a good semi-regular changing feature for the front page. This is also when stylesheets truly came to the Moonspeaker, as well as several other background changes most folks never see related to the advent of xml and the later html4 standard. All this before w3c began to advocate for digital rights management and it still seemed like the flash player was a safe thing to use. The layout is still a bit crufty and busy, especially the grey-scale bar at the bottom and the unnecessary extra labels, but this is definitely much better than before. This is also the "geneva" era in terms of the default site font. I admit, there is no real excuse for inflicting that font on those with no defence from it at the time, except for the fact that I hated helvetica then and hate it now. I'm not too much of a fan of arial either. I found geneva less objectionable due to its greater apparent width, which kept the text from looking too cramped. Stylesheets were in but the guidelines for web typography were still in their infancy, making it hard to keep things looking decent in multiple screen sizes. That said, I had already given up designing for 480 by 680 pixel screen sizes by then because I literally hadn't seen any for years except on machines that couldn't handle images when browsing the web anyway.
Unfortunately 2009 was also a crisis year in terms of editing and managing the Moonspeaker's content. By early april of that year, a new version of textwrangler had been released, and I had updated without a second thought, only to discover a terrifying bug had manifested. Somehow changes were being lost when I switched between documents, which destroyed an entire evening's work. My notes from the time briefly describe a bug report from barebones software which came down to a memory management problem triggered by a similarly timed macos system update. So began my sojourn in the world of emacs, a world as fascinating as it was infuriating. Text editors are like operating systems, programs that people once engaged in holy wars on the web about. Things are a bit calmer now. This particular screen grab is from aquamacs rather than plain emacs, though either runs under macos easily enough, and plain emacs is already installed anyway. However, emacs keybindings are a nightmare for a longtime mac keyboard touch typist, though I still have altering them in plain emacs in mind as an exercise for learning more about the lisp programming language. That is pretty much the only difference between the two versions of emacs, as you can see for yourself if you like. It is well worth digging around in emacs to see what's there. According to legend, there are hackers who open an emacs session and never use any other program for the day because it includes games, an email viewer, and a calendar application among many other things besides the basic text editor. It certainly reflects its origins in the time before the notion of a graphical user interface got loose. Nevertheless, as soon as the textwrangler bug was well and truly gone, I switched back and stuck with it until well into 2014. A quiet addition to the Moonspeaker in 2010 was an rss feed, something I had wanted for quite awhile but then had to spend far too much time reverse engineering and standards hunting before it was finally working consistently in 2012. The rss documentation is so frustrating that eventually I wrote up my own short guide on how to build a feed by hand. It's quite doable.
The 2013 era Moonspeaker included a light website redesign, including implementation of the by then existent web typography guidelines. Besides providing a couple of alternative stylesheets, I also changed font philosophies and stopped trying to enforce specific fonts. It's a fool's errand on one hand, but also understandable that so many web designers got caught up on it. The nature of the web as a medium took considerable time to get used to, especially the lack of control that its more fluid nature dictates. A nice feature that became truly standard fare across web browsers at this time is the tooltip or title, which could now be applied to any html tag. It's a small way of adding interactivity to a website, yet it is also a wonderfully non-intrusive one. It's a shame the capture didn't manage to catch the image with the featured thoughtpiece, but then again, the front page of the Moonspeaker today is evidence enough that images have become more prominent on the site. As a regular feature, thoughtpieces actually predate this view of the Moonspeaker, although they were rather thin on the ground for awhile. These days I try to make sure a new one appears each week. The Random Site of the Week feature became a bit of a bear to deal with for awhile due to the impact of blogging and the associated linking practices between blogs. This corresponded with some interesting changes in search engine philosophy led by google, shifting to favouring similar results and pushing different but related items much further down the results. 2013 is also the first year a fancy spammer tried to convince me to post ads on the Moonspeaker, which was quite funny.
Things got much busier in 2014, so much so that finally textwrangler couldn't handle all of the tasks that I wanted to throw at it, and I opted to switch to bbedit. It includes an ftp client as well as built in access to version managers like perforce and git, so bbedit has already more than recouped its price in its utility for many other tasks besides building the Moonspeaker. A few milestones happened in 2016. Although the rss feed reached 100 entries in 2013, it hit 200 in 2016 and is well on its way to more than 300 this year. Thoughtpieces finally got their own folder as part of the implementation of site search via duckduckgo. Several new essays and stories appeared in 2016 as well. So far 2017 has started at a reasonable clip, with the number of external links now firmly above 1 200, 100 random sites although not all have featured on the front page yet, and enough illustrated pieces to bring the image total over 100. For fun I keep an eye on how big the site is in terms of what media it can fit on. It still fits easily on the third smallest thumbdrives that used to cost a fortune only a few years ago, which held only 128 megabytes. Almost hard to imagine in these days of fairly affordable terabyte drives. No doubt there are plenty of readers who are unimpressed by those numbers. After all, the Moonspeaker has been around for well over a decade. Still, not many websites have been around this long with only one designer and content writer. Even multi-editor blogs are challenged to keep going past their fourth year, so fourteen seems a real accomplishment.
To this day the most linked to parts of the Moonspeaker are the draft of Amazon Nation and various items from the Fiction section. As might be expected, there is plenty of varying opinion about the quality of the material in Amazon Nation, whether the different chapters are really about the greek Goddesses or the roman ones and so forth. Would that "graeco-roman" mythology were that easy to tease apart. The Sappho project is beginning to draw some link attention, and it will be interesting to see how that develops as more material goes up for it. As a graduate student I have learnt to appreciate introductions and prefaces far more than I used to, and it has been interesting to observe the misinterpretations any of us can be prey to online when we come upon a web page via search engine and therefore somewhat out of context. That strikes me as something that can work to the detriment of any website of the Moonspeaker's type, since it has a range of diverse writing and translating on it.
As to what's next for the Moonspeaker, well there are plenty of possibilities. One is a sort of commemorative ebook with a "best of" selection of thoughtpieces, stories, and essays. Completing the annotation of Sappho's poems is a big goal for this year alongside finally getting a long delayed podcast into regular production. The site continues to be an effective way for me to explore both the limitations and the benefits of web pages as a medium. Things will remain as opinionated as ever, although another major website design is not in the works any time soon. I'm disinclined to completely upheave things simply because they've been the same for a few years, though a stylesheet overhaul is not unlikely. In any case, here's hoping for at least another fourteen years.