We posted the new AllPeers screenshots last Thursday, but apparently it took a few days for them to percolate through the blogosphere. As it happens, I set off on Monday for a week of vacationing in London, and left my laptop at home under the assumption that “nothing happens in the week between Christmas and New Year’s Day.” I’ve therefore been cut off from the web world for the past few days, an experience which is both liberating and slightly disconcerting for an internet junkie like myself. As a result, I was totally oblivious to the fact that people appear to be interested, after all, in the idea of adding peer-to-peer capabilities to Firefox.
This changed yesterday when I turned my cellphone back on after Mike Leigh’s excellent new play “Two Thousand Years” at the National Theatre. Slashdot had picked up the story, a text message from Cedric informed me, and as a result our web server had been battered senseless, again. I dare say we’ll get it right this time and find a configuration capable of withstanding a sudden massive spike in visitor volume. In the meantime, apologies to anyone who tried to access the site in the past twelve hours and wasn’t able to.
Some of the comments to the screenshots post merit answers, and I’ll get to that next week when I’m back in the office. After all, there’s much more to AllPeers than peer-to-peer. I was also planning to write about Digg, the latest challenger to Slashdot’s supremacy, riffing off Alex Bosworth’s excellent analysis. Now’s that we’ve been dugg, I have a bunch more to say about this. And finally, I may even post some predictions for 2006. After all, everyone else is doing it (Blake Ross’s satirical take is particularly entertaining). Right now I’m sitting in the Institute of Directors, and the frigid weather in London has turned what should be a cosy web-enabled nook into an internet igloo. For the sake of my frozen fingers, I’ll sign off.
I love it when someone defends a viewpoint I subscribe to, and does so convincingly and eloquently that I don’t feel I need to add anything. Such is the case with Stowe Boyd’s treatise “Traitors in our Midst: Web 2.0 Antihype”. The whole point of the term “Web 2.0″ is that the trends we are experiencing, while real, are too subtle and complex to describe in a more constrained way. Believe me, I spent weeks racking my brain for an appropriately pithy encapsulation.
Stowe writes, “I am not prepared to pen a magisterial debunking of the Web 2.0 antihype that is growing, but I am committed to chip away at it, day by day.” Hear, hear! Pass me a pickaxe and count me in.
We’ve been working incredibly hard on AllPeers 2.0, and over the past couple of weeks it’s really started to take form. Like Ernest and Julio Gallo, we will sell (well, give away) no revolutionary browser extension before its time, so we’re planning an extensive internal testing phase and private beta before we go live. In the meantime, we’ve got a new website online sporting our suitably bombastic new tagline: “The best thing to happen to Firefox… since Firefox.” Of course, hubris ain’t always bad, and it’s my fervent hope that this assertion will be borne out over the weeks, months and years to come.
We’ve also posted the first screenshots of the new Firefox-based AllPeers, and they’re stunning. You can see them all here with some marketing yada yada, but if you prefer a slightly more technical angle, I’ll say a few words about each one here as well (click on thumbnail to enlarge):
The heart of AllPeers is the Navigator, a sidebar that provides rapid access to all the information managed by AllPeers. Since AllPeers is by nature an extensible platform, the Navigator needs to be extensible as well. That’s why you see the buttons at the bottom for changing the view (”My Peers” and “My Media”). In this case, we’re looking at the “My Peers” view, where I can see all my contacts directly in the web browser, along with presence information so you can see who’s online. This view is currently used for sharing media files, and in the future we’ll be adding a presence API so that it can be hooked into other presence-aware software like instant messaging and VOIP clients.
 Here’s a media view. In this case, we’re looking at files that I’ve imported into the AllPeers database. Or, more precisely, I’ve imported metadata about these files (including thumbnails). I can browse these media files in Firefox alongside standard web content. We’re planning many features to make the integration between local and web content even more smooth, including support for importing anything from a webpage into your local datastore using drag-and-drop.
 When I want to share one or more of my local media files, I just pop open the Share taskpanel, which is a form that appears on the righthand side of the browser. (We hate dialog boxes so all our forms use the taskpanel approach.). Just click on the groups or individual users you want to share with and off you go. They will be alerted directly in their browser that you have new photos, videos, podcasts or whatever for them to download at their leisure.
I should also mention that AllPeers includes a complete BitTorrent client that you can use inside Firefox. We’ll be posting more information about this (and screenshots) at a later date.
I’m writing this using Performancing for Firefox, a WYSIWYG blogging extension. In essence, this is Flock for Firefox, something that I predicted would appear sooner or later. My overall impression is very positive. As with Flock, setup was easy and painless, involving the entry of my blog URL, user name and password. It then downloaded a list of all my existing posts and categories, so it’s much quicker and easier to update entries without dealing with all the latency of the web-based Wordpress interface.
The editor appears in the bottom half of the browser window in a resizable pane, so I can continue to surf around (including moving between tabs) while I blog. It’s easy to drag and drop hyperlinks and images into the editor, something that’s a real pain to do with a standard textarea. You can edit in the WYWIWYG view or move to a source view where all the HTML tags are visible. And, as a bonus, it allows you to create notes that are stored locally and can be searched. This may end up being the killer feature for me, since I often have thoughts and ideas but no convenient place to put them.
I do have a few nitpicks. Hyperlinks dragged into the editor retain their existing formatting, whereas I am far more likely to want them to take on the formatting of the surrounding text. They are also enclosed in a <cite> tag, and there’s no way to remove it without moving to the source view.
I’ve been using Scribe for months, so I’m used to pressing Ctrl-S incessantly to save my blog posts and avoid losing work. Performancing lets you save using Ctrl-S, but it saves the text as a note which remains even after the post is published. It also prompts you to confirm that you want to overwrite the existing version, since it doesn’t appear to know that you are saving a new version of the same text. I’d much prefer for blog posts to be saved as drafts somewhere (as Wordpress does natively) and deleted automatically when they are published. And Ctrl-S has to work silently (i.e. without confirmation) in the standard case.
The formatting of the actual HTML published to the blog is a bit opaque. Both the title and body of my last post had a bunch of whitespace preceding them, though this only showed up when I viewed the source (i.e. it’s not visible on the blog itself). And the tags and text run together in a way that is clearly made for computer consumption only. I wouldn’t relish editing these posts manually (e.g. from an internet cafe without access to Performancing). It would be great if there were some kind of pretty-printing applied before publishing.
I also discovered a seeming bug when I forgot to set the category for my last post. I was able to view the post in Performancing’s history, but when I changed the categories and selected “Publish as Edit”, the blog was not updated.
In general, the editor still has some hiccups, and it’s too often necessary to have recourse to the source view to correct non-visible elements like the above-mentioned <cite> tag. Also, the source view doesn’t sync up with the WYWIWYG view when you skip over, so you have to scroll down manually. But in general, the editor represents a huge step forward compared to my previous approach of using a textarea souped up with Scribe and Julien Couvreur’s Textarea Resize script. Only time will tell whether this becomes my preferred blog authoring method, but it’s great to see that we’re making progress.
Julien threw me for a loop by pointing out, in response to my last post, that start.com does exactly what I was postulating that Google couldn’t do, enabling developers to place arbitrary widgets on their homepage without the need for centralized approval. (This also adds some weight to Mike’s comment.)
The key point is that start.com’s “gadgets” are written in DHTML and therefore run in the web browser’s sandbox, with no access to local data. This means that your homepage — your homepage! — can’t display information such as the number of unread messages in your email inbox, the number of contacts online in your favorite email client or the amount of free space on your hard disk.
If you believe in the model where all data and code, including things like email and instant messaging, are stored on some central server and exposed through a web API, then you might disagree with my last post. (Note, however, that there are some serious privacy and security issues here with respect to the company doing the aggregation.) I actually do see a real possibility of this happening in the very long term, but not until there is universal wireless connectivity and a couple of orders of magnitude more computing power/bandwidth available. It might well never happen. Until it does, I stand by my assertion that data aggregation needs to take place on the client.
Google’s new Homepage API is a perfect illustration of the weaknesses inherent in the server-centric web model. Note that I would have to submit my code to Google for approval just to modify my own homepage. This creates an unfortunate bottleneck where I have to wait some unspecified period of time and conform to potentially arbitrary criteria before I can enjoy the fruits of my labor. It also creates a perverse ownership model since I can’t host and distribute my software myself. This is a big deal for me, since I like to keep all my creations here on this blog. I would have no problem listing my homepage module (if I were to write one) in the Google directory, but primarily I want it to sit alongside my blog posts, essays, article and other software, creating a composite view of who I am and what I’m up to.
Contrast this with the client-centric approach used for Firefox and AllPeers extensions. Yes, Firefox has an official extensions site with similar controls to those used by Google. But I’m under no obligation to use their site. I can put my extensions on my blog, email or IM them to my friends, burn them on a CD and mail 10 copies to every AOL subscriber or whatever. This is a much more fertile environment to spur innovation. It’s great that Google has an open API, but they need an open process as well, and this can only work if data aggregation and manipulation occur on the client, not the server.
After spending a good 10 minutes creating my New York Times Print Fixer, I was crushed to find out that it was seemingly useless, since an option in Firefox can be used to achieve the same thing (i.e. preventing NYT printer-friendly articles from rendering in a miniscule font). I duly disabled the script and turned off “Shrink to Fit Page Width”. Inevitably, I then printed a long article the other day, only to discover once I was comfortably ensconced in a crosstown bus that the last two words of ever sentence had been cut off. I soon gave up on trying to make sense of the text and instead spent the rest of the trip cursing the day I had changed my Firefox print options.
The funny thing is that, just as I was about to turn the print fixer back on, I decided on a lark to visit the NYT site and see if they hadn’t acted on my feedback of last week and resolved the problem. And guess what? They have! I’m very impressed. It’s rare that a large organization of any type reacts to the suggestion of a single customer this quickly. Looks like the print fixer is heading to the Greasemonkey graveyard.
I remember sitting in a Manhattan apartment a few years ago with an old college friend and his buddies, all of them alumni of the Wharton School at the University of Pennsylvania. When you study engineering at Penn, as I did, you tend to stereotype Whartonites as money-obsessed cynics, and the conversation that evening certainly bore out this prejudice. Specifically, our host, between shots of a very tasty iced Russian vodka, explained how the best way to get rich quick was to found a company, build it up over a couple of years and sell out for a few million bucks. As an entrepreneur who was trying to, you know, change the world, I was taken aback by this callous and unrealistic attitude.
So here we are a few cycles down the road, in the midst of what is increasingly looking like a brand new tech bubble, and once again the conversation has returned to “flipping” startups for a fast profit. (Irrelevant note: how appropriate that a Microsoftie uses an opaque GUID for his permalinks.) And strangely enough, I’m finding myself agreeing with those who defend this approach. If you can build a team, technology and/or user base that is valuable to a larger player, quicker or better than they can, you’ve created value and you deserve to be rewarded for your efforts. Not everyone has to create the next big public company in order to be considered a successful entrepreneur.
At the same time, I have a lot of sympathy for the naysayers. The objectionable part of the flipper philosophy is the assumption that it is somehow simpler or easier to make money by cobbling together a buzzword-laden product and selling the farm than by pulling down a salary in someone else’s company. The reality is that building a company, whether aimed at a quick sale or total world domination, is a huge and exhausting effort. As many have said before, you’d better love what you’re doing or you’ll never make it. What’s more, your odds of success are pretty slim, so you’d better factor in that risk if you decide to take the startup route. Yes, you could toil for three years and then sell your creation for $10-20 million to a big internet player, pocketing $2-5 million in the process. But if your chances of doing so are only 5%, you might be better off working for $100-150K/year as a software developer or manager. And if you don’t have that kind of earning potential, you may not have a resume that a Microsoft or Google would shell out millions for.
So if you’re buying into a new philosophy of corporate structure where innovation is partially “outsourced” to the startuposphere (my worst coinage ever?), your visions of flipping might be wholly rationale, and even admirable, as long as you are aware of the strenuous and perilious journey you’ve embarked upon. If you’re setting up a tech startup as an easy route to filthy riches, consider playing the lottery instead.
The web world’s latest unconfirmed rumor is that Google is planning to buy perennial third-place web browser vendor Opera. Everyone is citing Pierre Chappaz’s French-language blog and various execrable machine translations, so let me try my hand at a proper translation:
According to a normally very well-informed source, Google is on the verge of buying the Opera web browser. Although this information should be taken with a grain of salt for the time being, such a move on the part of Google would be a plausible response should Microsoft decide to integrate its own search engine into Internet Explorer more effectively than is presently the case. Microsoft’s search engine is actually in Explorer already, but hidden in the address bar where no one can find it, so it ends up being used only to return results if an address is entered incorrectly. For IE 7, Microsoft might take the lead of Firefox, which features more ergonomic integration of Google’s engine. In this case, what could Google do to avoid losing a very significant portion of searches in one fell swoop? It could launch its own browser… whence the Opera hypothesis.
This would make a whole lot more sense if Google weren’t so obviously in bed with Firefox already. I’m somewhat at a loss on this one, but here are some possible theories:
1) Google wants to grab as much of the non-Microsoft market as possible, so it sees an Opera acquisition as complementary to its Firefox activities.
2) Google wants to make inroads into the mobile space, where Opera is strong (a possibility raised by Gary Price).
3) Google is planning to create a Firopera browser that encompasses the best of Firefox and Opera.
4) Google is flailing and increasingly all over the map.
5) The rumor is false.
The only one of the above without critical weaknesses is the last. It’s hard to see how Google could motivate the troops to support Firefox so actively if it owned a competing browser. Firefox has its own mobile version, Minimo. It’s still in its infancy, but it would make more sense for Google to concentrate on that if they wanted to improve their support for mobile devices. Merging Firefox and Opera to get the best of both worlds is an appealing idea, but almost unimaginable from a software development perspective. And while Google’s recent activity has been a bit random, it’s hard to believe that a major strategic acquisition would take place without some sort of viable rationale.
So I can only conclude that the rumor, though intriguing, is unfounded.
Yahoo’s recent acquisition of social-bookmarking site del.icio.us gives me the perfect excuse to write about a topic that’s been brewing in my brain for a few weeks. I started thinking about the evolving roles of venture capitalists, large technology firms and startups after reading a Paul Graham essay, “The Venture Capital Squeeze”. Paul’s thesis is that startups are increasingly acting as outsourced R&D shops for bigger companies, taking advantage of the agility and innovative nature that are characteristic of small teams. VCs are thus facing more competition from Google and its ilk, who are keen to snap up cool startups before they are sullied by outside investors.
His solution is for VC funds, who are still swimming in cash, to allow a certain proportion of their investment to go straight into the founders’ pockets. This makes them more competitive since it tempers the stark choice that often faces entrepreneurs: get moderately wealthy immediately by selling out or accept funding in the hope of becoming truly stinking rich, but at much higher risk.
I agree wholeheartedly with Paul that traditional corporate structures are changing rapidly, in the technology space and elsewhere, and the idea that startups can and should be created with the explicit goal of selling their intellectual property and team to a larger player is certainly borne out by the del.icio.us acquisition. Del.icio.us is the quintessential Web 2.0 success story: a cool innovative idea and a healthy and fiercely loyal user community, but no mainstream awareness or business model. At the rumored purchase price of $25-30 million, it is probably a good buy for Yahoo. After all, this sum is a drop in the bucket for Yahoo, so even a free service without built-in revenue streams can be valuable if it improves overall user experience and thus revenue generation on other Yahoo properties.
At the same time, I can’t get myself to accept the idea that entrepreneurs should reject VC money unless they are allowed to cash out partially. First of all, the experience of del.icio.us is an excellent illustration of why raising VC money can be good for company founders even if their goal is a short-term strategic sale. It’s probable that the del.icio.us folks made out better financially for having accepted VC and angel investment in April. Yes, this diluted their stake in the company, but having strong financial backing gives you huge leverage when negotiating with a potential acquirer. Had they raised only enough angel money to keep their servers ticking over, they probably would have experienced a much more stressful few months and still ended up with less money.
And I simply can’t bring myself to believe that cash payouts by VCs wouldn’t reduce the incentive to go for a big win. The lure of riches is a large part of what keeps us working 14 hour days, sacrificing our social lives, boring our friends and family and all the other things that entrepreneurs endure. I’m certainly in favor of founders earning a decent wage once VC funding is secured, but a large cash payout is a bad idea. And add to this the potential for fraud. In the last bubble, companies were set up purely for the purpose of enjoying access to large sums of money. These perverse incentives would get that much stronger if unscrupulous “entrepreneurs” could put investment cash directly in their pockets.
Apparently I’m not the only one missing Scribe after upgrading to Firefox 1.5. So I tweaked the install.rdf file in my already souped-up version, and it seems to work fine.
I’ve taken the liberty of making my modded version available for download. If you are the creator of the Scribe extension and you’re mad at me for expropriating your work, please drop me a line. I’m sure we’ll find an amicable solution. In my defense, I sent you a mail months ago and you never answered.
In Slate magazine, Adam L. Penenberg argues that music companies should use automatic pricing methods to maximize their online revenues, instead of going with the 99-cents-for-everything approach used by iTunes. Funny thing is, Joel Spolsky believes that the record labels are avoiding variable pricing because it creates second-class citizens among their offerings. On the surface, I tend to agree more with Adam. Like many others, no doubt, I generally gauge the quality of wine in a restaurant mainly according to price. But this doesn’t mean that I always order the most expensive bottle. Cheaper wines may not be as good but they’re, well, cheaper, and at the end of the day I’m more concerned about value for money than absolute quality. I don’t see why the same principle wouldn’t apply to music (and more generally media) sales.
That said, pricing solely according to sales volume strikes me as overly simplistic. This may work well for back-catalog items that were fully amortized years ago. But what if I have a Belorussian Balalaika Blues Band and I know that the potential market for my new album is only about 10,000 units? I might be able to make a fair amount of money by charging a premium price, since Belorussian Balalaika Blues fans, though few and far between, are probably keen to snap up whatever Belorussian Balalaika Blues they can get their hands on. It would be most unkind to price down this kind of long tail speciality item just because it appeals only to a narrow niche.
I do believe that the idea of dynamic pricing is brilliant and could potentially revolutionize the media market. Pricing of digital goods is a black art, at best, and I don’t see why we couldn’t seek assistance in something akin to the yield management techniques that budget airlines use to maximize revenues. I’m not a good enough mathematician to work out the details, but perhaps the system could finetune pricing by clustering media files according to consumption patterns. An older recording that could benefit from a lower price would likely be purchased at a slow but fairly steady rate, whereas a new work with a specific niche audience would sell most of its copies when it is first released. This type of distinction could be used to adapt pricing appropriately for these and other cases. The system could also fiddle the prices randomly and measure the effect on sales in order to expand the volume of statistical data that it uses to make its decisions. It could even vary prices based on geography and other factors. I’d be curious to know whether any research has been performed in this area, so maybe it’s time I put Google Scholar through its paces.
I’ve been irked for a while by the fact that printer-friendly versions of New York Times articles print using a teeny-weeny font that I can barely read. Naturally this is one of those issues that invariably bugs me after the fact, so I never took the trouble to look into what was actually going on.
I finally took the plunge the other day, and after some fiddling with the CSS stylesheet associated with the printer-friendly pages, I convinced myself that the culprit is some sort of Firefox bug. The articles, after all, print correctly in IE, and none of the behavior that I observed in Firefox (in particular, the fact that style rules associated with the document footer seem to affect the whole page) struck me as particularly rational. I may be wrong about this, since I haven’t seen this commented on elsewhere. Or perhaps I’m the only myopic Firefox user out there. In any case, I’ve now filed a bug so I guess we’ll find out whether I’m missing something obvious.
In the meantime, I did what any self-respecting Firefox hacker would do and greasemonkeyed the problem into oblivion. Turns out that removing the useless footer entirely did the trick, and required only a two-line script. If you’re having the same problem, check it out: New York Times: Print Fixer.
Update: So it turns out that this is an issue with the “shrink to fit” print option, which appears to be the default setting (see the Bugzilla bug referenced above for more information). It can be changed in the Page Setup dialog box (File menu). So the utility of my Greasemonkey script seems to be limited. I wonder if other people have encountered this problem and, if so, whether there is a way to improve the default behavior.
Update: It looks like the NYT has fixed the problem!
|
|
|
