Marein reviews fan-made Uru Ages in anticipation of open sourcing. Parts 1, 2, 3.
March 4th, 2010
February 3rd, 2010
Paul ‘chip’ Querna: Facebook & Open Source: Community is just as important as the Code
When you create an open source project, you gain almost nothing but a PR hit if there isn’t a community built around it. [..]
Just look at the massive community that has exploded around Apache Lucene and Apache Hadoop — Yahoo could of kept this infrastructure project internal, and sure, it might of fulfilled their original goals, but they wouldn’t of ever received the thousands of external contributions, which has turned the Lucene/Hadoop world into one of the most diverse and thriving open source communities of late, giving Yahoo a thousand times return on their investment in Hadoop.
Paul discusses the importance of letting company-created open source projects thrive by fostering a vivid community, and encouraging their use in competitors’ products against what might be considered common business sense.
Consider the contrast between two of Apple best-known open source projects: the earliest, Darwin, never took off the way Jobs had promised (and perhaps hoped for) back in 1999. Projects such as OpenDarwin (which eventually shut down), PureDarwin, GNU-Darwin have always lacked proper leadership and interest and suffered under opaque, unclear, apparently inconsistent policies on Apple’s part.
On the other hand, there’s WebKit: once given a proper website with a public blog, issue tracker and repository and allowing external reviewers, its success exploded — and in spite of the fact that many of Apple’s competitors now use it, Apple, too, reaps rewards. The sum of the contributions simply exceeds the initial loss of control and propriety by a wide margin.
Perhaps this kind of success is only possible with a limited amount of projects, but as a company, you have to be willing to take that risk.
January 23rd, 2010
This week, YouTube announced its HTML5 beta, based on an earlier demo at Google I/O 2009. Vimeo swiftly followed suit.
Compared to their Flash-based predecessors, these new video players have various downsides: no full-screen mode, no embedding on third-party sites and, at least for YouTube’s, no play/pause using the space bar. Other than that, this is great: it’s great because we no longer rely on a proprietary mechanism of Adobe’s that was originally designed to display animated vector graphics, not provide a UI and codec to play movies; it’s great because a standards committee (and an experiment-friendly sidekick) have created this technology, and will hopefully drive it forward; it’s great because, in fact, this has benefits in performance (far lower CPU usage; higher likelihood of GPU acceleration), accessibility and other areas.
And it works, right now, everywhere, without limitations.
Ah, who am I kidding.
It in fact works in the newest versions of Safari and Chrome (including Google Chrome Frame for IE), and, if you’re on Linux, Opera. It doesn’t in IE, and — here’s the sad, perhaps surprising part — not in Firefox either.
You may be misled into thinking that the HTML working group standardizing a <video> tag implies them also agreeing on the formats used by such videos, mandating one or more video and audio codecs and container formats to support. In fact, you’d be right that the original proposal looked exactly like that, requiring Ogg Media as the container, Ogg Theora for video and Ogg Vorbis for audio.
The big benefit of Xiph.org’s Ogg-branded series of codecs is their liberal approach — no patents, open specification, no cost, BSD or BSD-style licensing. In other words: pretty much anyone gets to use them if they want. Contrast that to the increasingly common H.264 video codec, which costs money to license and has patents. You’d think the choice is pretty clear: for the consumer and the developer, Ogg Anything means bliss and sanity.
It’s not quite as clear-cut. Apple and Nokia objected to HTML5 requiring browsers to implement Ogg Theora, citing difficulty of hardware acceleration (plenty of chips that specialize in H.264 decoding already exist). Apple added that, even though it is claimed Ogg Theora doesn’t have any patents, there’s no guarantee. Google doesn’t want to switch YouTube to Theora, in part because Flash already supports H.264 (so that’s what much of YouTube’s existing library uses), and because the latter achieves better quality at the same bitrate.
As such, the requirement was dropped from a later draft version of HTML5. Safari 4 ended up shipping supporting whatever codecs your installed QuickTime comes with (which includes H.264, but doesn’t include Ogg Theora unless you install it yourself
), Firefox and Opera with only Ogg Theora, and Google Chrome with both. I said Opera only supports Ogg Theory; as I understand it, Opera on Linux is a special case: it behaves similarly to Safari in that it supports whatever GStreamer happens to have (including, where available, H.264), whereas on other platforms, they instead bundle Ogg Theora.
So here we are: a <video> tag and plenty of confusion as to what it means to the end user. Will my browser play this? Will it not?
Controversial quality comparisons aside, one cannot deny H.264 is very widespread. Apple has been pushing it big time, Blu-Ray mandates it (and the late HD-DVD allowed it), and Adobe Flash has supported it since version 9. For pretty much anything “HD”, it has become the preferred format; the Windows Media-derived VC-1 never quite took off the same way, and Windows 7 even finally ships with it. If smartphones, portable video players, etc. do video, chances are they do H.264. The same simply cannot be said for Ogg Theora.
But perhaps ubiquity isn’t the point?
If you ask Robert O’Callahan, it’s not. Licensing a patent-encumbered, non-free codec is a difficult matter, but that’s not even the big problem. The alternative of having the operating system’s media framework provide the codecs (as is the case with Safari/QuickTime and Opera/GStreamer) doesn’t appeal to him for a number of reasons, including a support nightmare and code maintenance headaches, but most of all this:
Even if a volunteer produces a patch [to make Firefox use Windows's extensible media framework DirectShow for video codec support], I would not want to ship it in Firefox in the near future; let me try to explain why.
Probably most important: we want to focus our energy on promoting open unencumbered codecs at this time.
That was half a year ago. He has reiterated it since:
Users just want video to work. You Mozilla people are such idealists! Yes, that is the reason for Mozilla to exist.
As far as “existing” goes, this may be right. Were it not for idealists, Mozilla may not have survived as long as it did without a successful product. Firefox 1.0 shipped November 2004; Mozilla was open-sourced March 1998. There’s a variety of reasons it took so long (including the questionable decision to rewrite), but there’s little disagreement that idealism helped keep the dream alive.
But the article is less about Mozilla and more about Firefox, and I’d wager to say people in this context read “to exist” as more as “to have succeeded” than “to have been created in the first place”. And in that, I couldn’t disagree more.
“Free software idealism” may have played a big role in those initial six years of not much visibly going on: alpha geeks enjoyed tinkering with the bits and pieces of code, and tried various approaches (particularly Seamonkey), and eventually settled for something far simpler (which evolved from m/b through Phoenix and Firebird into Firefox). That wasn’t what brought Firefox to the masses, though. The masses don’t care about the license. They don’t care to take a look at code. They do care that it’s free as in gratis, that it works, and that it works better than what they were used to. And what they were used to was awful, because it was IE 6.
Firefox 1.0 was the fortunate (and this is part of what significantly accelerated development those days) situation of IE 6 having a deservedly horrible situation and no proper competition, aside from alternatives nobody cared about, and ones even fewer people cared about. A combination of beta geeks1 — those folks who fix their friends’/neighbors’/relatives’ computers — and the media made Firefox a huge success, and they had IE 6’s suckiness (particularly in the realm of security, or lack thereof), Microsoft’s refusal to provide new versions, and Firefox including cool features with it as good arguments.
In short, Firefox was a free and for the most part far superior alternative, and people had their beta geek friend to help them set it up. That is why it succeeded. I consider it foolish to believe that free software idealism pushed Firefox to the masses. It merely helped ignite its beginnings.
And therefore, we must ask ourselves which is more important: more idealism that may or may not make a difference another six years from now, or a solution right now that users will be pleased with?
I’ll take Pragmatic Choices For Consumers for $500.
January 5th, 2010
As judged by anecdotes, as well as by going through the various demo videos (yes, this means I haven’t held one in my hand, so my knowledge is limited):
Things I’m jealous of:
Things I’m not so jealous of:
Android has been developing at a rapid pace, which puts some wonderful pressure on Apple to hopefully make a nice leap with iPhone OS 4 and beyond, but it’s nowhere near the point where I’d even consider switching.
September 6th, 2009
We’ll start with the very end:
Daniel Eran Dilger is the author of “Snow Leopard Server (Developer Reference),” a new book from Wiley available now for pre-order at a special price from Amazon.
I don’t find it unreasonable to expect someone who writes such a book to strive for a certain level of accuracy. Having said that, let’s go:
More importantly, Apple is providing its users with additional options that benefit both Mac users and the open source community.
What additional options for the open source community does Snow Leopard provide in the PIM area? A quote later on provides a clue what Dilger apparently thinks Apple has added:
Because Apple makes its money almost exclusively from selling hardware, it has opened up its own Snow Leopard Server applications, Address Book Server and iCal Server, as open source Darwin servers that can be compiled to run on Linux.
10.5 Leopard introduced the Apache-licensed Darwin Calendar Server, a subset of Leopard Server’s iCal Server, and this has been continuously updated (Snow Leopard Server ships with version 2). But while Snow Leopard Server ships with the new Address Book Server, there doesn’t appear to be any open source project for that. Marketing-wise, the iCal Server page mentions:
To further the widespread adoption and deployment of these standards, Apple has made the complete source code for iCal Server 2 available through the macosforge.org website.
No such thing for Address Book Server. Maybe they’re planning on it; maybe they’ve even said they are — but so far, this looks quite untrue. Dilger goes on:
That means Apple is essentially giving away both the client (to Mac users) and the servers (to the community) in order to encourage the use of open standards in messaging and collaboration.
No, the clients (Mail, Address Book and iCal) are most certainly commercial, closed-source software. Of the servers, all three are commercial and closed-source, although a subset of one is available in an open-source fashion. Which, by the way, is great on Apple’s part — but let’s not deny that a good configuration interface adds plenty of value, and Apple does not provide that for free (or otherwise openly).
Next up, Dilger compares Apple’s trifecta of client apps to Outlook, with rather bold claims:
Integrated support for Exchange beginning with last year’s iPhone 2.0 means Apple’s mobile platform simply doesn’t need an Outlook client. Now Snow Leopard can also get by without Entourage/Outlook, thanks to new and improved baked-in support for Exchange in Mail, Address Book and iCal.
Microsoft has responded with the announcement that it will now be delivering a real (but still scaled back) version of Outlook for the Mac again
Now, unlike many, I’ve always been a fan of the separation into three apps. But even in 10.6, they are a far cry from Outlook being “simply not needed” or possibly to “get by without”. Public folders, anyone?
With Snow Leopard and the iPhone each now providing their own client layer for accessing Exchange Server, Apple can now offer its users alternative access to other server products as well, from its own MobileMe and Snow Leopard Server offerings to web services from Google and Yahoo. This effectively turns Microsoft from a direct seller into a wholesaler that has to deal with Apple as a middleman retailer.
I’m sure this made some vague sense when it was written. It doesn’t when it’s read. The entire section goes on about Sears, CompUSA, Netscape, IE, IIS and off-shore wind energy. Actually, that last one was a lie. But a discussion of PIM client/server solutions this is not. He could have discussed Netscape’s brief ill-fated journey into the groupware market, but he didn’t. Instead, he’s talking about Microsoft’s evilness, implying a dominant position IIS has never had (”Microsoft first took control of the client with Internet Explorer and then began tying its IE client to its own IIS on the server side with features that gave companies reasons to buy all of their server software from Microsoft.”), and then switches over to everyone’s savior Apple with their open sourcing of Address Book Server, which hasn’t in fact happened. Finally, Snow Leopard Server apparently includes a “Push Notification Server”, which Apple knows so much about, nine out of the top ten results from Google are all articles of his, or links to them. So let’s skip this entire part.
Apple’s support for Exchange and its promotion of its own Exchange alternatives are two sides of the same coin, in the sense that they use the same technologies.
Well, this certainly is exciting news for Microsoft, who didn’t even know until this point that their very own Exchange Server has support for CalDAV and CardDAV built right in. (To be fair, it does for IMAP and LDAP, although it’s typically disabled.) Wouldn’t you love it if your developers don’t even have to build features, your marketing doesn’t even have to promote them, and yet you get to offer them?
Apple built its support for Exchange using WebDAV
No…
, the open specification that Microsoft supports on Exchange Server as a way to deliver messages to mobile clients.
…and no.
While Exchange Server has support for WebDAV, and WebDAV is very much an open specification, it’s such a broadly-specified protocol for file transfer and versioning over HTTP that it isn’t intended for mail, contacts, calendars, etc. in particular, so Microsoft has had to layer plenty of proprietary additions on top of it.1 Yes, it uses WebDAV. No, that’s not all there is to it. It’s about as vague a claim as calling XML or CSV a format. For transferring files, WebDAV is a sufficient specification; for storing mails, contacts, calendar events, notes and more, including a ton of metadata, it’s incomplete — by design.
Apple did not license Microsoft’s Windows-only “Exchange Active Sync” software; it merely licensed the rights to implement a compatible EAS conduit with Exchange. Apple owns the Snow Leopard software that talks to Exchange.
This may be, but given that it’s in the same paragraph, I’m skeptical, and also question the relevance. It seems a poor and unnecessary attempt at making Apple look independent. Maybe they didn’t pay a licensing fee; instead, they had to pay their developers to develop client code of their own. So what?
The client applications Apple has upgraded in Snow Leopard to connect to Exchange, including Mail, Address Book, and iCal, also use WebDAV to talk to Apple’s own Snow Leopard Server applications.
The latter part is correct insofar as that CalDAV and CardDAV are extensions to WebDAV for calendars and contacts, respectively. They’re entirely incorrect for Mail (there is no WebDAV-based mail specification aside from Exchange Server’s proprietary method), as well as for Exchange.
This all leads to an entirely wrong conclusion:
This effort to support everything from integrated client software owned by Apple makes Snow Leopard’s support for Exchange of use to everyone, even if they don’t use Exchange. The client work Apple has invested in making Macs Exchange-friendly also improves the features available via MobileMe, Snow Leopard Server, and even some other third party services such as those from Google and Yahoo.
On top of incorrectly believing that Snow Leopard interfaces with Exchange Server through WebDAV, Dilger apparently goes even further that, since CalDAV, CardDAV and his imaginary MailDAV2 are built on top of WebDAV and Exchange uses WebDAV as well, Apple is saving duplicate effort. That would be great. It’s also entirely off. First, Snow Leopard communicates with Exchange Server through the much newer, SOAP-based Exchange Web Services protocol. That’s why it requires 2007 Service Pack 1 Update Rollup 4; this entire interface is lacking in 2003 (and in the original 2007 release). Second, even if they were to use WebDAV, this wouldn’t help them much at all.
Imagine two CSV files, one with the columns Surname, Name and Birthday, and another with the columns Last name, First name and Phone number. Superficially to the human eye, they both clearly contain contacts. To the computer, they’re entirely different formats. One column is missing from each other’s format, and while two out of three columns have the same contents, they’re differently named. You’d have to write a converter to make them match. You have the same situation with different XML formats3, and with two different takes at implementing, say, calendars on top of WebDAV. And given that Microsoft is moving away from WebDAV, citing lack of efficiency, they probably won’t implement CalDAV any day now.
App Store? Really? What does that have to do with anything?
The success of the iPhone App Store has benefited both developers and users by establishing a competitive market based on meritocracy. Snow Leopard’s support for Exchange, because it opens up equal access to alternative competition, similarly creates an iPhone-like market for desktop messaging services ranked by merit, not the vendor’s current market position.
Yes, vendors like, say, Microsoft. Also, software companies located in Redmond, Washington state. Can someone explain to me how an interface to a proprietary PIM protocol “creates a market ranked by merit”?
This will provide Snow Leopard users with not just the ability to talk to corporate Exchange Servers, but also the ability to access Apple’s own offerings and other third party services.
I’ll get right around to implementing Exchange Web Services in my own groupware. Surely the specification is somewhere on ietf.org. Oh, wait.