Paul Graham continues his popular series of "How To Get Rich Quick on the Internet. And Fast!" essays with his latest entry, "Be Good", in which he describes some startup attributes common among a sampling of successful internet businesses.
One such pattern, observes Paul, is that his selected success stories started with no real revenue plans beyond "get big and flip it".
No subscriptions. No advertisements. No pop-ups. No interstitials.
They started more like a charity than a real business, just bringing good to their userbase with nothing asked or expected in return. At least at first.
This isn't a new position for Mr. Graham. He has long advocated the idea that you just need to worry about getting the eyeballs and you can figure out how to make money from them later. Or better still — reading between the lines — you can let the sucker you flip the thing to worry about the gritty details of how to monetize it, suffering the consequences if the userbase burns it to the ground in defiance of any revenue generating scheme.
This isn't a surprising position for Paul to support. It is entirely aligned with his micro-VC organizations' business model, which is to take young, time- and energy-rich grads, fresh out of the college mill, and bankroll them with a small investment (which they need because they won't be earning anything from their Internet baby anytime soon), and then cash in when/if they manage to flip it to a sucker that still buys into the many eyeballs model (a strict "No Returns" policy in effect.)
Given the small investment, only a percentage of YC Combinator's `fundees' need to hit the jackpot for the strategy to succeed, at least for Paul. He's playing house odds in this startup casino.
Paul provides some examples to demonstrate his position that the charity-that'll-make-you-rich approach is the winning strategy: Google and Craigslist. Incontestably successful companies, and most would be over the moon to experience a fraction of their success.
Let's take a closer look at these examples, and see how relevant they are to Paul's central theme.
Craig Newmark, the founder of Craigslist, started compiling a list of upcoming local events in 1995, publishing it to subscribers via a listserver. Later, after the list was well established and had a healthy subscriber base, he started publishing entries to a website, adding functionality to allow users to email directly to categorized lists.
Today Craigslist is an internet superstar, constantly ranking among the top 50 websites worldwide. It pulls in impressive revenue numbers through a model that Craig himself describes in this entry on Yahoo! Answers.
Craigslist is a rare survivor among thousands, tens of thousands, or perhaps hundreds of thousands of exceedingly similar lists/classified upstarts, most of them run ad-free and for free (sites were often ad free as a simple side-effect of the barriers to entry to hosting ads pre-2000. At that time it wasn't as simple as signing up for an AdSense account).
Craigslist had the perfect, rare combination of the real-world personal connections of the founder, an ideal starting locality, userbase, and a progressive evolution that allowed it to build enough momentum that eventually the network effect took over, and you'd use craigslist because everyone uses craigslist, at least in some markets.
What You Can Learn From Craigslist: Craigslist is an extreme anomaly. Holding it up as an example of a path to follow is fundamentally akin to analyzing the number picking "strategy" of the latest lottery winner. It's also an excellent example of how corrupting, and falsely compelling, a survivorship bias can be. Countless sites have followed a close to identical path, failing miserably.
Google hit the scene during the portal craze. This was a period when every other search engine, failing to sufficiently profit off of search alone, started merging with a gangly bunch of dance partners. Excite hooking up with @Home, for example. To "leverage the synergy", each quickly morphed into a "destination for all things" portal, fattening up their content until they featured a landing page absolutely packed from margin to margin with text, news, stock quotes, images, comics, horoscopes, etc.
Ads were almost invisible in this era of visual pollution, and were seldom the problem. Many users were using dial-up, so the content inflation wasn't just aesthetically deplorable, but it also made the search process a slow, unpleasant experience.
Sergey and Larry had been working on some algorithms for internet search during this period, and with the dot com boom peaking they managed to pull together an impressive million dollars in financing before even launching their beta website.
When they did finally get something online, they did the absolute minimal amount possible. Later they described the utter simplicity of the first version as a function of their lack of HTML knowledge: It was the best they could do, or cared to do, at the time.
They had differentiated themselves, however inadvertently, and it worked brilliantly. They had copious long-term financing (not "flip it to someone else" financing) before even launching, so they had no need to worry about making money immediately, using the website as a technology demo that would conceivably allow them to sell search technology and services to third parties and businesses.
While the quality of the search results got the Google buzz started, the dial-up bandwidth-friendly simplicity of their offering really won people over. Yet it was a simplicity that came primarily because the company only really had one product — search — and couldn't link to hundreds of other provided services.
When everyone else went heavyweight, the minimalism of Google got it a lot of attention among technology trend makers. That exposure on sites like Slashdot — amplified when the community learned that Google ran Linux — got them their next $25 million in financing, and the rest is history.
With their advertising initiatives Google took exactly the same approach, and when everyone else was using pop-up, high-bandwidth, obnoxious ads, Google zagged and had text ads. As others have adopted Google's text ad approach, Google has started adding in animated and full graphics ads to their docket.
What You Can Learn From Google: If you have an algorithm or technology that is sufficiently impressive enough to get a million dollars in financing before you've even left the drawing board, maybe you can take lessons from Google beyond "differentiating from the competition is good" (which is fairly obvious advice.)
The vast majority of "go big of go home" web ventures will fail. It isn't a meritocracy. Luck has a lot to do with it. There is little you can learn from the success stories unless you also learn from the failures, yet aside from the huge flameouts (which had to have enough success to even be notable), most failures fizzle out and disappear without a trace.
From the opposite angle, many "grow revenue from day one" websites have succeeded admirably (I just gave a local babysitter directory $39 for 3 months of lookups), albeit not without the "lottery ticket" quick payoff that a miniscule percentage of the winner-take-all players yield.
It's a risky and disingenuous proposition to build a website, baiting a community, on one model — the "Good" and charitable model — and then switch the userbase once the founder's numbers get drawn. It's a scummy behavior to engage in, much less evangelize. It also might have the opposite effect than intended: I might not care whether a site like a social link voting site has a revenue model, but when I was considering photo sites I immediately discarded those that followed the Paul Graham business model, considering the risks too high: Either it would eventually be forced to flail about, obnoxiously trying different approaches at making it pay, or it would fold with a "Sorry We Got Bored Our Numbers Didn't Come Up" notice one fateful day. Instead I went for one with a sustainable business model, and haven't been unhappy with my decision (even if it did cost me a bit per year for all of the features).
Back in the late 1990s, when XML was still in its formative years (a state some would argue continues to this day), XML Spy was a very welcome entrant to the developer tools market, bringing intuitive, GUI-based schema and basic transformation authoring and validation to the developer’s desktop.
While some were productive and happy with just the W3C specs and a copy of emacs, many of us only used XML intermittently, building an export, import or transformation that simply worked, promptly forgetting all of the nuances of DTDs versus XSDs versus XDRs, or the quickly changing XSL(T) specifications.
It was a great step forward in the uptake and quality of XML utilization to have such an easy to use, up-to-date tool.
At the time XML Spy was basically shareware, offering a fully featured 30-day trial, at worst popping up the occasional “please register me!” exhortation.
Many just registered it: it was an easy sell at $54 a user, less if you bought multiple copies. That’s almost disposable money, and was an easy pitch to most managers. It was easy enough saying “let’s get a copy for everyone in the group. Even for the guy in the cube near the washroom, anti-XML rage bursting from his trembling lips in a spray of spittle and phlegm.”
Time goes on and we all moved to different projects, divisions and companies, often with long gaps needing little or no in-depth XML. When those instances came up, we’d try to find an old licensed copy, or would download the latest trial, using yet another toss away email address for the validation.
And XML Spy just kept getting more expensive. The company grew and grew (note that the domain on the original archive.org link above actually expired, and now sits in the hands of a domain ad purveyor), and the dollar signs in their dreams had them imagining, apparently, of a day when millions of information workers sat toiling their days away in the pure awesomeness of XML Spy. In emacs-esque form, it had grown more and more functionality, even if many users never used it for anything more than creating and validating schemas and transformations.
By late 2000, the price of XML Spy had inflated to $149 a user. By the end of the next year it hit $399 a user. By late 2006 it was up to $499 a user (at some point dropping the space between XML and Spy, becoming XMLSpy).
As I write this it’s up to $539 a user.
Maybe XMLSpy is developed in a poorly insulated aircraft hangar in Siberia, and thus is strongly impacted by the price of oil?
A ten-fold jump in price in about 8 years seems excessive. What was once a wonderfully priced utility is now a considerably expensive development ecosystem. What was once an easy purchase (at one workplace I just paid for it myself rather than deal with the annoyance of a requisition form) is now a difficult to justify expenditure, requiring vendor comparisons, and negotiations with middle managers. When the money handlers are convinced, often it’s just for partial coverage of the development team.
You end up with the “XML guy”, rather than having a team appropriately equipped with a uniform set of tools.
Of course, clearly my complaints are off base. Altova obviously did appropriate research, and they determined that there really are people and groups who’ll happily pay more for an XML editor than they paid for their entire Visual Studio suite.
But come on, Altova – bring back a, err, “Semi-Professional” version – something with XML schema and transformation authoring and validation and nothing more. No grand vision where your product is the center point of a developer’s existence. Put a reasonable price tag on it – like $59 – and I’m sure you’ll get a lot of sales where right now you get none. I realize you probably have lots of big buildings with expensive lights, and layers and layers of bureaucracy to finance, but don’t do it all on the back of a simple little XML utility.
One of the most referenced papers in software development has to be Dijkstra's seminal paper titled "Goto Statement Considered Harmful".
Dijkstra didn't actually author the title, but instead it was the creation of an editor en route to being printed in an ACM publication. It was changed from its original title of "A case against the goto statement".
While the core essence of the essay is indeed that the goto statement can be harmful, Djikstra wasn't making an absolute statement (as is commonly claimed, and which is an absolutism tendency of far too many in this industry), but instead was commenting on the abuse of goto that was occurring in the industry, calling for a sober evaluation of where it is appropriate, but more importantly where it is not.
Nonetheless, the meme was created and has been reused and abused in innumerable Considered Harmful declarations since.
A month or so back the development webosphere was awash with references to Scott Hanselman's excellent blog, all excitedly linking (rel="titillating"?) to his piece titled "The Weekly Source Code 13 - Fibonacci Edition". This was particularly common in the .NET community, with many linkers describing it as an elucidating example of the many advantages of .NET 3.5 / C# 3.0.
I perused the entry, always eager to absorb that sort of information, but found it less than perfect. I withheld critical comment, hoping it would all just blow away.
Then this morning I opened up Visual Studio and happened to notice a link to his entry on the Start page.
Maybe it's been there for a while (the last date is pretty old) and I just didn't notice it before, but the title used on the Start page pushed me over the edge, coercing me to comment.
There are several issues I have with Scott's Fibonacci entry.
First, the C# 2.0 (henceforth I'm dumping the subversion precision on the language versions) version is oddly dumbed down: C# 2 also has ternary comparisons, and it even has anonymous functions (including closure functionality). Yet the demonstrations given contrast the simplest possible C# 2 implementation with the most obtuse C# 3 example.
Basically the only novel difference with the C# 3 example is that it uses a lambda, though of course it would be an absolutely terrible thing to use a lambda for.
It's not a very good example of the implementation differences between the versions, which is the claim made by the Visual Studio start page, and was the description often used during the dissemination of this piece.
I like C# 3, but this isn't a good demonstration of any advantage of the language.
Worse yet, the only place you'll ever see recursion used to calculate Fibonacci numbers is in "Recursion for Dummies" type examples. To understand why that is, consider Scott's C# 3 example, which he leads into with the statement "Now, here's a great way using C# 3.0".
Here's a logarithmic-scaled chart of the number of function calls necessary to calculate Fibonacci numbers in the C# 3 example Scott gave.
Obviously it gets unusable pretty quickly. Try calculating the 90th Fibonacci number using recursive algorithms...
In the same way that Goto can be harmful, the use of recursion is often a sign of badness, and this is no exception. Epic inefficiency is used instead of the obviously simple approach.
long CalcFibonacciNumber(long n)
{ long current = 1, previous = 0, swapholder; while (n-- > 1) { swapholder = previous; previous = current; current += swapholder; } return current; }
(Ignoring mathematical shortcuts)
A lot of readers will be rolling their eyes right about now, muttering something along the lines of "Awww, come on...you didn't seriously think anyone thought that recursion was a good way to calculate Fibonacci numbers, did you? This is beginner's stuff, and no one really thinks that's the right way to do it!"
I'm optimistic about the profession, so no, I didn't really think it was a serious example (though I do think it nonetheless deserves some serious warnings to ensure no one becomes misled).
WARNING: The Code Contained In This Example Will Rot Your Brain. Never Do Something Like This In Real Life. Don't Let Peers See You Looking At Code Like This. Suspend All Critical Thought While Reading This Piece.
Instead it's a sample of "here's a demonstration of how to do something absolutely terrible — almost felony worthy — in a variety of programming languages....".
This is still a serious problem.
The example given is so very wrong — even if it is what's used in Recursion for Dummies books — that it makes it close to impossible to focus on the actual point being made, even if it had used comparable features of each language to demonstrate how the same task could be accomplished in each.
It reminds me of many early web service tutorials and advocacy pieces: Many used absurd examples like "a web service to add two numbers" (and amazing variations such as subtract two numbers, multiply two numbers, divide two numbers, compute the Log10 of a number, and so on. You get my point — things for which a web service would be entirely unsuited).
Stop it!
Stop with the ridiculous no-one-would-(or rather should)-ever-do-it-this-way examples. It completely undermines the value of the examples.
Surely there are realistic examples that would be more appropriate for demonstrating the advantages of lambdas (recursion {is recursion}; [goto {is recursion}], so there isn't much enlightenment provided there). How about "how to build a rudimentary regular expression parser in a variety of languages", or for a web service "pulling weather data from a remote weather station".
Something that a developer isn't going to have to slog through with their brain fighting them on every line, demanding an explanation for the terrible design or algorithm they're supposed to accept at face value.
I decided to take the new SunSpider benchmarks for a spin, generating the pretty graphs found down below. Benchmarks are always entertaining, and it was enjoyable comparing the numbers yielded under various conditions (turning SpeedStep on and off [none of the benchmarks loaded the CPU long enough for it to bother raising up from its lowered power, 66% performance relaxation state], setting CPU affinity, running it on different PCs, trying different build options on my Firefox build, etc.)
"Why benchmark at all?" one might ask. Simple: If you find the right measures, the common wisdom goes, the inputs to the measure will improve as the various players work to improve the metrics.
Whether you’re measuring bugs per developer, lines-of-code, widgets per hour…whatever: Start measuring it and invariably it’ll start moving in the desired direction, whether this actually serves your end goal or not. Often such initiatives come at the cost of the unmeasured, but over time it adapts and starts serving as a beneficial feedback.
During the summer in my late teens I worked
on an assembly line building car parts (pieces that played some sort of role in
the air conditioning system – basically widgets): Put a little bracket in a metal
cylinder, add a circle of fiberglass, inject some desiccant beads using a machine,
add another fiberglass circle and another metal bracket to hold it all in place
and then put it in another machine that squashed another cylinder onto the top.
Then I sent it down the line to the welder.
Atop my machine sat a little counter that monitored my progress, carefully recording every piece assembled. While this was a less advanced era — being the prehistoric early 90s and all — and I had to manually transfer the final count to my timecard for submission, every worker was kept somewhat honest by the metrics submitted by the other workers on the line.
Clearly I couldn’t have done 2000 parts in a day if the people before me and behind me in line only reported 1000, for instance, and vice versa.
Coupled with continuous, careful QA tests and random inspections (performed by people who had their own metrics to work towards), this struck me as an excellent system because it was difficult or impossible to game, and the onerous checks ensured that it didn’t come at the expense of quality.
It certainly worked wonders on me, as I wiled away the endless summer days performing the most awesomely brainlessness of tasks by competing with my own personal productivity “records”, endlessly trying to push out more quality parts per hour day after day.
I was there and had nothing better to do, and that little counter sat looking down on me, mocking me. It dared me to do just a couple more pieces per hour, and I willingly complied.
Somewhere a paper pusher and cackling middle manager would sum up the part counts and rub their hands together in giddy glee, eager for my zombie-state quest for worth to pad their bonus cheques.
It’s good I was a summer employee, because my pace didn’t make me friends on the line.
Test Driven Development tries to create a similar spirit of metrics, giving you a goal to strive for as you build out your product. It’s a comforting bit of feedback when all of the TDD tests come back with green checkmarks. The more tests you create, the higher the absolute count of passes you can brag about when the product sails through with flying colors, easily passing 497 of the 497 tests.
Performance benchmarks serve the same purpose for the performance and efficiency domain.
Consider the initial hardware-accelerating video cards for Windows. Early on they seemed to have little or no purpose, and were almost abstract to users. Then benchmarks started appearing, giving the manufacturers something to strive towards while also providing end users with an easy way to compare and choose amongst the options. “Card {A} can only do 10,000,000 accelerated rectangles per second, while card {B} can do 12,000,000. Clearly we need to get card {B} for our rectangle displaying needs.”
Of course some vendors started gaming the metrics in various creative ways (see Joel's excellent essay on poorly thought out metrics). Several created products that actually recognized running benchmarks in hardware, “optimizing” (by any means possible, including simply discarding many of the benchmark commands, knowing the end user will never notice if every second rectangle or rendered text of the millions per second isn’t being rendered). Worse, the benchmarks were so atrociously artificial, bearing little similarity to actual everyday use that the direction of progress was to optimize the performance of benchmarks, often to the detriment of everyday use.
Eventually the benchmarks matured, getting better and more realistic, and the gaming was prevented or embarrassingly heralded, and it became a hugely beneficial tool in the march forward in the field. Various games have served the benchmarking role, the Doom and then Quake series being the most influential.
In the browser market, the growing interactivity of the web and the renewed competition amongst the big competitors has seen a flurry of benchmarks being widely discussed and debated, stereotyping each of the browsers into performance ghettoes. “Firefox is sooooo slow….” “IE is garbage. Opera is super speedy!”
Having some real tests is of obvious benefit to “set the record straight”, not to mention that it provides a carrot for the competitors to chase. Exactly that happened with me a while back when I came across a string concatenation benchmark, so I went in and streamlined the piece of Firefox code specifically impacted by that benchmark. My change in place, Firefox indeed did much better on that specific benchmark, though the real-world benefit was negligible.
In many ways the various web benchmarks available reminds me of the early accelerated video card benchmarks: Crude, having little or no correlation with the pain points of real-world use, and opportune for gaming and false evangelism.
Which brings me to the recently released SunSpider benchmark (which is a credible contender for the widely coveted “most poorly chosen project name” award: It’s bad enough that an Apple project uses “Sun” in their product name, but it's thrice as bad when it’s a project related to JavaScript – JavaScript being another nominee).
SunSpider is very easy to run and gives quick feedback, so quite a few charts and graphs have been sprouting on blogs across the land.
JavaScript/DOM performance is a huge concern right now, as web applications are growing in richness by leaps and bounds, so there is definitely a need to be filled.
Will SunSpider be what we've all been looking for?
Here’s just such a graph, charting the stacked benchmark runtime for the current tier-1 browsers for Windows.
Benchmarks were performed on a 4GB, Q6600 quad-core Core 2 processor machine running Vista x64. Firefox 3 was built from the current CVS (as of this morning). The Y-axis represents milliseconds.
Such a benchmark provides immediate feedback regarding the biggest bang for the buck optimization, at least in regards to improving the runtime of this particular benchmark. For IE 7 it is pretty clear that the benchmark killer is the bizarre and repeated use of string concatenation throughout the benchmark tests, particularly evident in the string-base64 and string-validate benchmark.
After approximately 20 seconds (okay, maybe 22 seconds) "optimizing" the base64 and validate tests to use the extremely common Array push/toString idiom that is used on pretty much any page that does more than the most trivial of string operations (my changes were rash and very simplistic, though if I were motivated — if this were production code — I could do a much better job with it), the performance had changed rather dramatically, as seen in the following graph (scroll up and down for dramatic flair).
It's late and I'm tired, but I'd guarantee that I could dramatically decrease the remaining largest test -- string tagcloud -- but I think the point is proven.
Some will naturally draw from this the presumption that I'm just an Internet Explorer 7 fan, desperately manipulating the benchmark to best fit the strengths and avoid the weaknesses of my favourite browser.
My browser of choice is Firefox. Not only do I not find the featureset of Internet Explorer 7 uncompelling and anemic compared to a naked copy of Firefox (not evening considering the enormous functionality offered by add-ins, such as the extraordinary Firebug), I find the performance of Microsoft's offering to be atrocious on real-world websites.
I don't like Internet Explorer on technical grounds, and I like it even less given the concerning conflict of interest it represents.
Perhaps I'm just bearing a grudge.
We're currently implementing a very rich, advanced web application, and one thing that we've found, in case after case, is that in real-world situations with extensive DOM manipulation and production JavaScript, Internet Explorer stumbles and groans under the load, while competing browsers complete the task with gusto (just rendering a dynamically loaded complex table takes 20x or more on IE than in Firefox 2. The disparity grows greater with Firefox 3). It's to the point that I can't help but wonder if Microsoft is trying to undermine the whole web thing intentionally, hoping to encourage the middle-grounders to hoard to the boards proclaiming the deficiencies of web apps, manipulated into begging for some XAML goodness.
So if I wasn't looking to defend IE7, what was my point?
Maybe the motivations of the team behind this benchmark were noble, and they weren't blinded into naturally biasing the benchmark towards their own project, but I can't help but see this benchmark as an entirely artificial, naive, unrealistic benchmark that adds little to the benchmarking landscape. A cursory glance through the benchmark sees bizarre oddities that would never appear in real-world code, and a variety of implementation choices that are questionable for a benchmark (for instance test/sample data is often constructed within the timed scope of the benchmark in the SunSpider tests, as if a production website needs to create 4000 random email addresses and ZIP codes, for instance. Normally such data is constructed outside of the timed loop, for obvious reasons).
The lack of weighting, the lack of realistic test scenarios... I'm just not convinced that it holds much utility (though I do like the way they have the "driver", and the elegant and clean client-side way they aggregate the test values, and do the same for comparison. The framework is a great foundation) for cross-browser comparison. I can see use in analyzing performance differences for a single browser (the results turning Firebug on and off, for instance, were very surprizing), just not as a valid comparison between different browsers.
Just as I dramatically changed the IE results in less than a minute of code changing, I'd guarantee that I could do the same with the other outliers (in particular the longer Firefox tests).
I'm still waiting for a good, real-world benchmark. Something that simulates sites like Digg, Slashdot, Facebook, interacting with them in a way that a real world user really would.
Some recent hardware upgrades left one of my PCs in a fragile state: While I had managed to avoid the time sucking irritation of an OS reinstall over generations of hardware changes, the latest hardware swaps -- or rather wholesale migrations -- pushed the crufty stack of refusing-to-vacate legacy drivers and confused minibuses over the edge.
It had become a fragile platform, with frequent crashes and unexplained behaviour.
Not surprising, given that the OS install had seen the platform go through a myriad of trasitions, including from a VIA chipset to an nforce chipset to a VIA chipset and then back to an nforce chipset; from nvidia to ATI to nvidia graphics.
Really it had dealt with change remarkably well, at least relative to the tradition of Windows being tightly bound to the installation hardware, with the slightest deviation demanding a reinstall.
Like many users, I'd been avoiding Vista. Actually that isn't entirely true: It wasn't that I was avoiding Vista, but rather that the benefits it offered just weren't compelling. Not compelling enough to reinstall a working system, or even to deal with the inevitable dev team just had to move stuff around to set their fingerprint search for traditional features.
When Flip 3D is a highlighted feature, you know that things can't be that great.
Nonetheless, given that a rebuild was in order anyways, and given that a professional in this field should generally ride the leading the edge of technology (though nothing like the days of old -- once upon a time we had to eagerly consume every Microsoft beta as a necessity for job security, whereas nowadays technologies like WPF are a complete and utter non-issue in the general tech industry), I figured I'd force myself to embrace Vista, warts and all.
I'd eat my industry's dog food.
To make things more interesting, I decided to go full bore and install the x64 version of Vista. I had a fully 64-bit capable processor, and despite only having 2GB of RAM (with two slots eagerly waiting to be populated), it would nonetheless allow me to start playing around with the extended x64 instructions and registers.
x64 is much, much more than just bigger pointers, so I wanted to play around with them in my favourite development environment (like many such transitions, it will take a few years before the benefits of x64 are fully realized, though in the .NET world the 64-bit runtime immediately takes advantage of the new features for platform neutral assemblies.)
The install and transition went very well. While I had to disable UAC quickly to make the experience marginally pleasant (I'd experienced Vista enough in dual-boots and virtual machines to quickly learn to despise this security "feature"), generally it just went well.
On the modern nforce4-based motherboard (featuring onboard sound and networking, all drivers installed by the single package of nvidia drivers), with a Q6600 quad-core processor and contemporary nvidia video card, everything simply worked after running a couple of driver packages. Even my dual-layer DVD burner just worked. All of the attached Canon printers, including the multifunction with faxing and scanning, auto-installed and immediately worked. Visual Studio 2005 worked great with the Vista patch, and while it's a 32-bit app it capably generates great 64-bit builds. Awesome!
Now I had IIS 7 to play around with, getting ready for the same in the server sphere. The integrated pipeline and web.config based configuration really is a wonderful step forward for the platform.
Several things didn't work. CameraWindow won't work with my Canon Digital Rebel XT, though by switching to PictBridge on the camera I found that Vista itself did a great job downloading and organizing the photos (and Windows Photo Gallery is a universe better than the terrible "Picture and Fax Viewer", not just for navigating and organizing photos, but moreso because the venerable XP viewer did a horrible job resizing photos for display at monitor resolutions, managing to make great shots look terrible). The bundled burning software that came with my DVD drive wouldn't install, but I was delighted to find that Vista actually had pretty decent burning capabilities built in, competently writing 9GB backups to my DL discs.
I hadn't really exploited the benefits of the 64-bit address space at this point, but then I booted up a copy of Battlefield 2 for some late night mental-diversion fragging.
Imagine my surprise when this game that ran very well on a vastly less powerful 1GB XP system now yielded frequent stutters while significant memory paging occurred. Despite having twice the RAM to draw from in the rebuilt machine (if you add in the system baseline, it really had about 2.7x as much RAM to draw from), somehow it was now hitting the ceiling on a machine that was almost completely dedicated to it.
So I grabbed 2GB more PC6400 DDR2 from a local retailer, plugged it in, and with 4GB everything is running pretty smoothly. Another $120 to upgrade the memory to have the same experience that I had with 1/4 the memory in XP.
Vista was hardly a surprise to me given that I'd been playing with it in various forms since the Longhorn days, and the technical feature list is entirely underwhelming compared to the early vaporware promises, however it isn't the travesty many are making it out to be. While it definitely needs work, it isn't going anywhere, and Microsoft isn't going to rollback to XP.
If you have modern, supported hardware, Vista x64 is a solid choice.
I recently opted to throw together my own blog software (after going through the standard Build or Buy analysis), expediting deployment as a means of forcing follow-thru. The goals of this micro-project were to improve the authoring and content management experience, to improve searchability of the content (without having to cast content out from the blog to a static form), and to improve the usability and navigation from the user's perspective (for instance the classic "date" navigator common on most blogs is something that I've opted to remove).
Despite having close to no time to allocate to this task, my tendency to over-engineer still showed through: The easiest option would have been a content-management system defined entirely in code (it's as easy for me to change and deploy code than it is to change templates and metadata), and of course to build it for a single author. Instead it supports many blogs through the same URLRedirector, blog aggregations (where a blog is a publication of a set of blogs, each with distinct authors) each using its own templates and configurations.
Which brings me to templates -- failing to find a decent Smarty-type templating system for .NET (basic ASPX is really a templating system, but I'm speaking more towards something that can enumerate sections, retrieving data based upon an object structure of relationships and containment).
So I had to build a basic templating system, yielding the templates that follow. The first for HTML output--
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8"></meta>
<title>{#blog.Title} {#docTitle}</title>
<link rel="stylesheet" type="text/css" media="screen, projection"
href="http://www.yafla.com/dforbes/style/css/blog.css"></link>
<script type="text/javascript" src="http://www.haloscan.com/load/dforbes"> </script>
</head>
<body>
<div class="clsHeader">
<div class="clsBlogHeader"><a href="{#blog.BaseUrl}">{#blog.Title}</a></div>
<div class="clsSubheader">{#blog.Description}</div>
</div>
<div class="clsBody">
{foreach $entry in $entries}
<div class="clsEntry">
<div class="clsDate">{#entry.EntryContent.PublishDateUTC|dddd, MMMM dd yyyy}</div>
<div class="clsTitle"><a href="{#entry.Permalink}">{#entry.EntryContent.EntryTitle}</a></div>
<div class="clsBody">{#entry.EntryContent.EntryContent}</div>
<div class="clsKeywords">{foreach $keyword in $entry.EntryKeywords}
<a href="{#blog.BaseUrl}{#keyword.KeywordText|escape}">{#keyword.KeywordText}</a> {/foreach}
</div>
<div class="clsPermalink">
<a href="javascript:HaloScan('{#entry.MappingId}');" target="_self">
<script type="text/javascript">postCount('{#entry.MappingId}'); </script></a>
<a href="{#entry.Permalink}">permalink</a>
</div>
{foreach $relatedentry in $entry.RelatedEntries}
{ifcond $LoopFirst = "True"}
<center>
<div class="clsRelatedEntries">
Related Entries
{/ifcond}
<div class="clsRelatedEntry">
<a href="{#relatedentry.Permalink}">{#relatedentry.EntryContent.EntryTitle}</a>
</div>
{ifcond $LoopLast = "True"}
</div>
</center>
{/ifcond}
{/foreach}
</div>
{/foreach}
<div class="clsAdBlock">
{#adBlockHorizontal}
</div>
<div class="clsNavigator">
<span class="clsNavigateEarlier">{#moveEarlier}</span><span class="clsNavigateLater">{#moveLater}
</span>
</div>
</div>
<br/>
<div class="clsAttribution">
<a href="mailto:{#entry.EntryContent.ContentAuthor.EmailAddress}">
{#entry.EntryContent.ContentAuthor.Name}
</a> -
{#entry.EntryContent.ContentAuthor.Description}
</div>
</body>
</html>
The next template is for RSS consumers--
<rss version="2.0">
<channel>
<title>{#blog.Title|escape}</title>
<link>{#blog.BaseUrl}</link>
<description>{#blog.Description|escape}</description>
<lastBuildDate>{#buildDate|r}</lastBuildDate>
<language>en-us</language>
{foreach $entry in $entries}
<item>
<title>{#entry.EntryContent.EntryTitle|escape}</title>
<link>{#entry.Permalink}</link>
<guid>{#entry.Permalink}</guid>
<pubDate>{#entry.EntryContent.PublishDateUTC|r}</pubDate>
<description><![CDATA[{#entry.EntryContent.EntryContent}]]></description>
</item>
{/foreach}
</channel>
</rss>
All in all, I think it works pretty good, and I can successfully run the W3C validations on the vast majority of generated pages and get the comforting green checkmark.
My original foray into the land of blogging was delayed while I stumbled towards the goal of building my own blogging software: like many software developers, I have a sometimes irrational desire to build it myself rather than admit “defeat” and use one of the many (and in the realm of blogging, there are many) available products.
I took a couple of stabs at building it myself originally, but due to another common foible – a tendency to over-engineer (I couldn’t simply write some blog software to post and publish my own thoughts. No…it had to be a full multi-author aggregation and collaboration suite, meaning that weeks went by while I mentally debated the database model for such a machination) – it just never seemed to get finished.
Other priorities always trumped it, and the little time I did allot towards this goal saw me solving absurd edge conditions.
I finally set a deadline for myself, and when I couldn’t find the time to finish anything before my marker (billable hours always came first), I went and bought a copy of Radio Userland and started publishing content the blog way.
That worked well enough for a while, but Radio Userland is a venerable publishing tool that is really showing its age. Authoring to it is a less than pleasant experience – which has been a huge contributor towards the dearth of content (it’s always a bit of a roll of the dice to see which characters it randomly replaces in posts, or which carefully authored HTML blocks it’s decided to mangle) – and simple tasks like cross-linking posts (e.g. a “related posts” sidebar to allow users to easily see follow-ups) was just far too manual to be worth the bother.
Now that I have a powerful, fully dedicated server, it’s also grossly under-featured for users, making the experience of consuming and navigating through the information far less usable than it should be.
So I’ve gone and built my own blogging software, this time quickly bringing it to a sort of beta release.
Given that this is the venue with which I will publicize a ton of changes elsewhere on the site, I really considered this a roadblock on the critical path to the release of other web application functionality elsewhere on yafla.
With some focus, it took only a couple of hours this time, mostly accomplished while putting my toddler son to bed over the weekend. It was so ridiculously quick and easy that I kick myself for not having done it sooner.
I’m extremely pleased about the functionality built out (hey it isn't rocket science, and definitely falls within the realm of "trivial", but there's lots of little "gotchas" with software like this), though most of the kudos go towards .NET 2 and SQL Server 2005: A couple of tools that make short work of what would once have been an enormous task, bringing a robust, secure, high performance web application to a usable stage in less time than it takes to watch the Lord of the Rings trilogy.
Right now you’ll probably notice that – at this moment at least – the HTML version of the blog looks absolutely terribly. That is somewhat by design (or rather an intentional time compromise)…momentarily. I’m working on the template (it’s of course parameterized template driven), and wanted to force myself to follow through by deploying (perhaps prematurely).
So what are the features of the blog software?
Well, firstly I migrated 100% of the old content over (including metadata such as categorization), running it all through Tidy first to try to make it a little more XHTML legitimate. Using an identifier mapping structure, every single link to the legacy content still works (which was important to me: I didn’t want to give link followers the frustrating “We moved everything so have fun trying to find it” 404 experience).
Everything works via URL remapping, and for now I’ve set it to redirect from old links to the new links where possible. E.g. http://www.yafla.com/dforbes/categories/softwareDevelopment/2005/09/28.html redirects to http://www.yafla.com/dforbes/Clean_Code. All new entries Will follow that more transparent and obvious structure.
But the URLs aren’t limited to just single documents – All entries in June of 2006 can be accessed via http://www.yafla.com/dforbes/2006/06. Add in a category and you can refine further – http://www.yafla.com/dforbes/2006/06/.NET (or http://www.yafla.com/dforbes/.NET/2006/06. Whatever makes you happy).
Want that in RSS form? http://www.yafla.com/dforbes/2006/06/.NET/rss.xml. Add in the day if you wanted to refine further.
Of course, no longer are entries limited to the archaic “categories”. Now they’re basically keywords, so if you want to see the posts where I’ve abused categories and multi-tagged, take a look at
http://www.yafla.com/dforbes/.NET/SQL/Blogging/SoftwareDevelopment/Personal/IT/
Yikes!
So the tagging will be much more logical now that there aren’t broad categories, and given that anyone can filter content however they want (stick rss.xml on the end and you can get a feed of whatever you want).
There’s also search, though I’m not comfortable enough with the finality of the API to publish anything about that.
Entries now have versioning, given that I want to be more transparent with edits that I make (I’m endlessly doing minor corrections and improving wording, and for those who consider that deceptive there’ll be a little version history to see what changed and when, along with a label of why the change was made). All links are auto-parsed and logged, so every entry has a list of posts that link into it, making for much more elegant self follow-ups without resorting to post-editing some “UPDATE: See also…“ notes into old entries, and without resorting to the ugliness of trackbacks.
Extensive caching ensures that it’s still spritely and capable of handling peak loads with no fuss.
Oh, and the system supports many blogs by many authors, including publishing multiple authors into one system…so I still over-engineered, but in the end it was workable and I’m extremely happy with the core structure.
Great things lie ahead.