Older blog entries for pabs (starting at number 41)

Talk to Your Kids About YAML or Their Friends Will

This article nicely illustrates my beef with YAML; namely, the YAML grammar is far too complicated. Complexity in data serialization and exchange formats should be avoided, because it virtually guarantees subtle interoperability problems.

If you need to exchange structured data in a language-agnostic format, do me a favor and use JSON instead.

Syndicated 2009-04-15 16:38:00 from Pablotron: News

Joggle 0.1.0 Released: Jabber to Twitter Relay

I just released Joggle version 0.1.0. Joggle is a Jabber to Twitter relay; tweets show up as instant messages, and instant messages are posted as tweets.

Setting up Joggle is easy; all you need is Ruby, five minutes, and a spare Jabber account:

  # install joggle, create joggle directory
sudo gem install joggle
mkdir ~/.joggle

# create joggle config file 
# (replace joggle@example.com and abc123 with your spare jabber 
# account and password, respectively)
echo -e "jabber.user joggle@example.com\njabber.pass abc123" > ~/.joggle/joggle.cfg

# run joggle in the background
joggle --daemon --config ~/.joggle/joggle.cfg

Next, add the specified Jabber account (joggle@example.com, in the example above) to the buddy list in your Jabber client.

Finally, register your twitter username and password with Joggle by sending an instant message like this: .register TWITTER_USER TWITTER_PASS (replace TWITTER_USER and TWITTER_PASS with your Twitter username and password).

You can also share your Joggle installation with your friends; have them add the Jabber account to their buddy list and send a .register command.

See the README file for detailed installation instructions and a full list of configuration options.

Files:

Update: Comments for this site are still broken, but I've cross-posted this release announcement on Reddit, so feel free to post a comment over there.

Syndicated 2009-03-14 20:05:38 from Pablotron: News

ZipStream-PHP 0.2.1 Released

Version 0.2.1 of ZipStream-PHP is out. There is one change:

  • Generate correct "version needed to extract" header. This fixes extraction problems with WinZip 9.0.

Here are the relevant links:

Syndicated 2009-03-09 22:43:02 from Pablotron: News

I Have a Twitter Account

Despite my better judgement I now have a Twitter account. You can follow my wacky hijinks at http://twitter.com/pablotron.

I've also written a basic Jabber/Twitter relay called Joggle. I'm hoping to release it later this week. If you'd like to try Joggle now, you can grab it from the mercurial repository (you'll need Ruby, a spare Jabber account, and a few minutes to glance over the README file).

Syndicated 2009-02-22 06:27:29 from Pablotron: News

Gratuitous Hardware Imagery: New Firewall and Fille Server

I put together a new firewall and a new file server. The firewall is an ALIX 2d3 -- a tiny, low-power x86 SBC with onboard 3xLAN. The whole thing runs on 7-20V. Also, there's no heat sink ahe disk is compact flash, which means no moving parts!

The other machine is a new file server. It's replacing three older machines and a half-dozen Vservers. The hardware isn't as exotic as the ALIX, but I do have some pictures:

Syndicated 2009-02-22 05:57:19 from Pablotron: News

Techmeme Author Filter 0.2 (Greasemonkey Magic)

I don't read Techmeme and you shouldn't either. But if you do, then this Greasemonkey script for Sean might come in handy:

A brief README file is also available. Basically the script allows you to hide articles by author. You can configure the filters using the two buttons in the top-right corner of the page.

Syndicated 2009-02-22 05:26:35 from Pablotron: News

ZipStream-PHP 0.2.0 Released

Version 0.2.0 of ZipStream-PHP is out. The two changes are:

  • Generated archives work with the Windows XP "compressed folder" feature.
  • Dropped support for PHP 4.

Here are the relevant links:

As you can see from the link above, I'm using Redmine for bug and feature tracking. I've been using it for several months and it's fantastic.

Syndicated 2009-02-22 04:57:58 from Pablotron: News

PersistJS: Cross Browser Client-Side Persistent Storage Without Cookies

I just released PersistJS, a client-side JavaScript persistent storage library. Features include:

  • Small (9.3k minified, 3k gzipped)
  • Standalone: Does not need any additional browser plugins or JavaScript libraries to work on the vast majority of current browsers.
  • Consistent: Provides a consistent, opaque API, regardless of the browser.
  • Extensible: Custom backends can be added easily.
  • Backwards Compatible: Can fall back to flash or cookies if no client-side storage solution for the given browser is available.
  • Forwards Compatible: Supports the upcoming versions of Internet Explorer, Firefox, and Safari (Opera too, if you have Flash).
  • Unobtrusive: Capability testing rather than browser detection, so newer standards-compliant browsers will automatically be supported.

If you already know why this is awesome, then you can skip straight to the download. If you're scratching your head, then read on...

Why This is Awesome

Why use PersistJS? What's the problem with using cookies directly or simply requiring Flash?

Currently the only reliable cross-platform and cross-browser mechanism for storing data on the client side are cookies. Unfortunately, using cookies to store persistent data has several problems:

  • Size: Cookies are limited to about 4 kilobytes in size.
  • Bandwidth: Cookies are sent along with every HTTP transaction.
  • Complexity: Cookies are difficult to manipulate correctly.

Modern web browsers have addressed these issues by adding non-Cookie mechanisms for saving client-side persistent data. Each of these solutions are simpler to use than cookies, can store far more data, and are not transmitted along with HTTP requests. Unfortunately, each browser has addressed the problem in a different and incompatible way. There are currently 4 different client side persistent data solutions:

  • globalStorage: Firefox 2.0+, Internet Explorer 8
  • localStorage: development WebKit
  • openDatabase: Safari 3.1+
  • userdata behavior: Internet Explorer 5.5+

Some developers have attempted to address the client side storage issue with the following browser plugins:

  • Adobe Flash
  • Google Gears

The problem with relying on plugins, of course, is that users without the plugin installed miss out on the feature in question, and your application is dependent on software from a particular vendor. Google Gears, for example, is not widely deployed. Flash is, but it has problems of its own:

  • Many users block Flash or require a click in order to enable flash content; this makes Flash unsuitable as a transparent, client-side data store.
  • Flash is notoriously unreliable on newer 64-bit machines.
  • Some businesses block Flash content as a security measure.

Anyway, if we include Gears and Flash, that means there are no less than 6 incompatible solutions for storing client-side persistent data.

The most notable attempt at addressing this problem is probably Dojo Storage. Unfortunately, Dojo Storage does not support Internet Explorer without Flash, and it does not support Safari or other WebKit-based browsers at all (at least, not without Flash). Also, Dojo Storage is not standalone; it requires a several other Dojo components in order to operate.

PersistJS addresses all of the issues above. It currently supports persistent client-side storage through the following backends:

  • flash: Flash 8 persistent storage.
  • gears: Google Gears-based persistent storage.
  • localstorage: HTML5 draft storage.
  • whatwg_db: HTML5 draft database storage.
  • globalstorage: HTML5 draft storage (old spec).
  • ie: Internet Explorer userdata behaviors.
  • cookie: Cookie-based persistent storage.

Each backend exploses the exact same interface, which means you don't have to know or care which backend is being used.

Examples

Here are some simple examples of PersistJS in use:

  // create a new client-side persistent data store
var store = new Persist.Store('My Data Store');

// pretend data
var data = "pretend this is really long data that won't fit in a cookie";

// save data in store
store.set('saved_data', data);

That's all there is to creating a persistent store and adding some data to it. Fetching data back from the store uses a callback function (to support asyncronous backends), but it's still pretty simple to use:

  // get data back from store, and prompt user with it
store.get('saved_data', function(ok, val) {
  if (ok)
    alert('saved data = ' + val);
});

Removing data is pretty easy too:

  // remove data from store
store.remove('saved_data');

Although it isn't necessary, you can also get some information about the detected backend using the Persist.type and Persist.size attributes:

  // build alert message
var info = [
  'Backend: ',
  Persist.type || 'none',
  ', ',
  'Approximate Size Limit: ',
  (Persist.size < 0) ? 'unknown' : Persist.size 
].join('');

// prompt user with information
alert(info);

Finally, if you don't want a particular backend used under any circumstances, you can disable it using the Persist.remove() function:

  // never use cookies for persistent storage
Persist.remove('cookie');

Download

This is the initial release, so there are bound to be some bugs. PersistJS has been tested with FireFox 2.0, FireFox 3.0rc1, IE7, and Safari 3.1. The Gears and Flash backends work where Gears and Flash 8 are supported.

The most notable omission here is IE6; it should work, but I don't have IE6 handy at the moment (MultipleIEs is temporarily busted).

Commenting is still busted around here, so any comments should sent via email or posted on the Reddit thread.

Syndicated 2008-05-22 07:19:49 from Pablotron: News

Don't Use ExtJS

A couple of years ago I recommended YUI and Ext (formerly YUI-Ext). I've changed my mind. Don't use Ext at all. The Ext license has changed four times since its inception, and each time the license has become more restrictive.

History

Ext was originally created as an extension to YUI. It was BSD-licensed, just like YUI. YUI-Ext added several sorely-needed features to YUI. The most notable additions were a layout system, a Tree View widget and a Data Grid widget (YUI has since added each of these, although the YUI widgets are still less flexible than their Ext counterparts). Eventually support was added for jQuery and Prototype as well. The team dropped the "YUI-" prefix, and YUI-Ext became Ext.

Ext 1.0 was relicensed under the LGPL. Although switching from the BSD license to the LGPL is relatively innocuous, it is still significant because the LGPL is more restrictive than the BSD license.

Eventually the Ext team changed the license again. The new license was a custom license that granted conditional LGPL usage rights. Basically the LGPL usage clauses applied, but only if you weren't trying to develop a library or an Ext clone.

Confused? Yeah, me too. Here's the text from the old Ext "Open Source License":

Ext is also licensed under the terms of the Open Source LGPL 3.0 license. You may use our open source license if you:
  • Want to use Ext in an open source project that precludes using non-open source software
  • Plan to use Ext in a personal, educational or non-profit manner
  • Are using Ext in a commercial application that is not a software development library or toolkit, you will meet LGPL requirements and you do not wish to support the project

A lot of open source developers were understandably confused by this hybrid license. The biggest problem was that it wasn't clear whether this license was compatible with other Open Source licenses. It also wasn't clear whether Ext could be legally distributed with Open Source software, since the license only granted LGPL usage rights, and not LGPL distribution rights.

To address these complaints, the Ext team changed the license again: the latest version of Ext is licensed under the GPLv3. This latest change complicates things quite a bit for many users, as we'll see in the next section.

Problems

The GPL is far more restrictive than the BSD license and LGPL. It is rarely used for libraries, because the viral clause would effectively the library from being used for any non-GPL software. In fact, these problem were addressed over 16 years ago by creating the LGPL:

By 1990, it was becoming apparent that a less restrictive license would be strategically useful for some software libraries; when version 2 of the GPL (GPLv2) was released in June 1991, therefore, a second license - the Library General Public License (LGPL) was introduced at the same time and numbered with version 2 to show that both were complementary. The version numbers diverged in 1999 when version 2.1 of the LGPL was released, which renamed it the GNU Lesser General Public License to reflect its place in the GNU philosophy.

Who is affected by this change? In no particular order:

  • Extension Authors: Older Ext user extensions could be licensed as the author saw fit. This is no longer true for the latest version of Ext; new user extensions must be GPL-licensed, because the viral clause prohibits using Ext with non-GPL licenses.
  • Commercial Users: The previous licenses, even the questionable custom license, allowed Ext to be used in closed source commercial applications. This is no longer true for the latest version of Ext, because the viral clause prohibits using Ext with non-GPL commercial licenses.
  • Non-GPL Open Source Developers: The BSD and LGPL-licensed versions of Ext could be used with other non-GPL software. This is no longer true for the latest version of Ext, because the viral clause prohibits using Ext with non-GPL licenses.

Things get even murkier when you consider linking and distribution. Does generating a dynamic page count as linking to Ext? Does any public web application automatically count as distribution? What about applications which use Ext to and access a common APi, such as a SOAP endpoint or RSS feed?

These questions were posted in a thread about the license change on the ExtJS forums.
Here's how Jack Slocum, the primary Ext developer, responded:

If you are generating any markup or javascript code via the server in a page SPECIFICALLY designed for Ext, then that server code will have to be GPL as well.

For example:

  • Suppose you have an index.php that includes Ext JS. According to the FSF, in that case index.php would be also under GPL since it is using ext. Since it must be GPL, it's source must be distributed. Since it is GPL, the "viral" effect of GPL is now in effect and any thing that uses index.php (if anything) on the serverside would also fall under the GPL. (Note: Note this is a pretty gray area)
  • Suppose you are using server-side code to generate javascript that interacts with Ext JS. That code must also be GPL.

Like MySql and other GPL software the way to use GPL code without having to license under GPL is to not bundle or distribute the GPL code with your application. If you instead have the end user (developer?) download and install ext js on their own, they are then bound to the license and not you or your software.

For those seeking an FAQ, we have defined and explained some of the reasoning and license implications under these 2 pages:

http://extjs.com/products/license.php

http://extjs.com/company/dual.php

It's worth noting that the examples given at the beginning of this post are just my opinion and it is impossible for us to analyze everyone's usage and say whether or not someone "complies" with the GPL. That really is a task for an attorney or even someone with better knowledge of your application and how Ext JS is used.

In the end, we want Ext JS to be open source friendly and still have a good business model in place to grow. The old Ext License was not open source friendly and pretty much killed all options for use in open source projects. That wasn't our goal so we had to address it.

There are several problems with the statements above. The biggest one is that the original BSD license and subsequent LGPL license had none of this ambiguity.

In other words, the problem the Ext team is trying to fix is one they created themselves. If that wasn't bad enough, the solution actually hurts many Open Source developers far more than it helps.

In an attempt to clarify the situation for non-GPL Open Source developers, I posted several questions in the Ext license thread. I also created a post on Reddit about the license change and summarized my questions there:

As of page 8 of the thread on the license change I have yet to receive a response, simple or otherwise, to any of my comments:

There's this comment:

The new license prevents Open Source software that is using a license other than the GPL from using Ext. Applications which use popular Open Source licenses like the LGPL license, BSD license, MIT license, and the Artistic license would be required to either re-license under the GPL, carefully design their application to meet the requirements in your post, use an older LGPL-licensed version of Ext, or move to another library entirely.

And this one:

What about authors who which to provide their software under a license that is more permissive than the GPL, such as the MIT or BSD licenses?

And this one, which was directly in response to Jack Slocum, the primary author of ExtJS:

Hi Jack,

I can see how switching to a license without exceptions would make things simpler, but what about those of us who release Open Source software under non-GPL licenses such as the BSD, MIT, and Artistic licenses?

I've been an Ext user since its inception as YUI-Ext, but the fact that I cannot seem to get a straight answer to a simple question makes me wary and extremely skeptical.

John Resig, the author of jQuery and Processing.js, responded:

It's important to understand that OSS developers are not their target audience at all. I'm 100% certain that we'll never get a clear response. They're using 'open source' as a buzzword selling point to lure companies in, befuddle them with confusing viral licensing, and obligate them (through the obvious balking that the corporate lawyers will do) to get them to buy a full, corporate, license. It's very sneaky, quite disingenuous, and paints a bad picture for open source development as a whole.

It's been over three weeks since these this exchange on Reddit. None of my questions have been answered on the Ext license pages or in the 68-page license thread on the Ext forums.

Rationale

According to the Ext license page, Ext licensing is based on the principle of "Quid Pro Quo", or "something for something":

Dual Licensing is based on the principle of Quid Pro Quo - "something for something". In return for the advantages you realize from using an Ext product to create your application, we require that you do one of the following:

  • Contribute to the continued development of the product by purchasing commercial licenses from Ext. This option secures you the right to distribute your application under the license terms of your choice
  • Contribute to the Open Source community by placing your application under an Open Source license (e.g. GPL v3). This option secures all users the rights to obtain the application's full source code, modify it, and redistribute it.

The justification for using the GPLv3 instead of the LGPL is addressed on the Ext license FAQ page:

We considered once again releasing under straight LGPL but it was not an option as a business. We tried that with version 1.0 and found out quickly that it enabled others (e.g. large commercial entities) to take our work, wrap it up and sell it as their own. With no mention of us at all. We, as a business with a full time team of talented developers, can not exist under those circumstances. We would quickly become diluted and competing with ourselves.

The concern about others taking their work and selling it without attribution is particularly ironic, considering:

  • Ext only exists because a large corporation (Yahoo!) decided to share their hard work under a permissive open source license
  • Significant portions of Ext, including the Event handling code and reset.css, were copied wholesale from YUI
  • The Ext object system comes from YUI, which is based on Dean Edwards' JavaScript inheritance code
  • Many of the older Ext icons originally came from the Famfamfam Silk icon set

What do the projects above have in common? That's right, they are all in the public domain or available under extremely permissive Open Source licenses.

The Ext team is certainly entitled to license and sell their software any way they see fit. However, it is hypocritical and dishonest to complain about other people taking your work and selling it as their own when you take other peoples' work and either sell it as your own or relicense it under an extremely restrictive license.

It is tempting to attribute this entire fiasco to a simple misunderstanding on the part of Jack Slocum and the Ext team. Here's what I had to say on Reddit:

It is a bit disconcerting that Ext has such strong roots in existing Open Source software, and yet the project seems at best partially indifferent, and at worst, outright hostile to the Open Source community.

Unfortunately, according to John Resig, this isn't the first time that there have been problems with the Ext team:

We (the jQuery project) worked hard with them to try and fix bugs and add features for an ExtJS integration layer. They turned around and built their own, specialized, library (removing the need for any of our work) and then mutated the licensing into this bizzaro scheme that they have now. We can't, in good consciousness, even recommend their library anymore due to its very nature. On top of this they ended up hiring our lead evangelist to promote their work. I can't speak for everyone on the team but I feel quite frustrated and used.

They're providing a great disservice to the Open Source community in general. They consume with reckless abandon, it's impossible to even hope to borrow code from them, and they turn it all into a money-making machine. No aspect of that sits well with me.

Jack Slocum did respond to this comment on a separate blog. He also wrote a post on his blog. Neither adequately addresses John Ressig's main points or my questions from the Ext forums, so I won't bother quoting his mostly vacuous responses here.

Conclusion

To recap, the reasons I recommend against using Ext are:

  • A trend towards more restrictive licenses
  • Team is unwilling or unable to address licensing issues
  • Blatant disregard for other Open Source projects

Some suggestions for the Ext team:

  • Release Ext under the LGPLv3, BSD license, or GPLv3 with the special provision
  • Provide the Ext artwork and CSS under one of the licenses above, or one of the Creative Commons attribution licenses
  • Update the Ext license FAQ with detailed information about which Open Source licenses are compatible with Ext, including specific usage scenarios
  • Add a list of Open Source software which Ext has borrowed from to the license page and to the license information included in the download
  • Put older versions of Ext back on the download page

Finally, here are a list of Ext alternatives. None are as nice as Ext, but they are all available under permissive licenses and they each have an active and enthusiastic user community:

Comments are still broken at the moment. I've posted this article on Reddit, so feel free to comment there.

Syndicated 2008-05-15 07:26:16 from Pablotron: News

EasingJS, a JavaScript Easing Library

I just released EasingJS 0.1.1, a port Robert Penner's ActionScript Easing library to JavaScript. EasingJS allows you to easily generate smooth and stylish animation or color transitions. For some examples, check out the test page.

Syndicated 2008-05-14 22:25:49 from Pablotron: News

32 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!