Tag Archives: JavaScript

Synchronization Using Interval Tree Clocks in JavaScript

As a follow up to my previous post, I’ve implemented the Interval Tree Clock code in JavaScript with tests.  I’ve also begun a synchronization framework to go with it.

Github ITC in Javascript

The framework would be for synchronizing documents in full mesh mode — so peer to peer.

Everything has tests, so you can easily see the direction and progress by just reviewing the tests.

I stab the Synchronization with big knife

Synchronization With Interval Tree Clocks

Sync ProblemsI’ve been working with mobile devices for a long time, and inevitably the most painful piece of the development process is getting data to be consistent across all replicas.
For years, I’ve been trying to find a consistent means of taking care of this in a way which is OS and repository agnostic for all replicas. It isn’t 100% clear to me why this isn’t a solved problem, but I have a feeling there are several contributing factors:

  1. Internecine conflict between all relevant parties.
  2. Rapidly changing means and standards for data storage and transmission.
  3. Figuring out causal relationships between data on different replicas is really, really difficult.

It seems to me that number 1 and 2 having become somewhat better lately because of ubiquitous JavaScript.  I’m not saying it’s trivial, but you can make an app that works just about everywhere now if you write it in HTML and JavaScript.

When dealing with data, browser based apps are still likely to be a problem with large data sets and long periods without connectivity, but it might be worth exploring the possibilities again.

To this end, I’ve been looking at solving the causal problem with Interval Tree Clocks (ITCs) lately.  They are interesting in the way that licking battery terminals is interesting.  They are painfully tedious, but if you can stick with it, you may eventually power a solution (or be brain damaged).

For a long time, I think the standard way to handle the problem of causal relationships has been vector clocks, but they have well documented limitations around space usage which do not apply to Interval Tree Clocks.

Also, you can make pretty diagrams with ITCs.

ITC Node Diagram

So I’ve been trying to rewrite the ITC algorithm in C#.  This may seem ironic since I just told you that JavaScript seems to be one solution to some of the industry’s synchronization problems, but the reality is, I’m much better at exploring ideas with type safe code.

I’ve gotten most of the C# working, and I’ve created tests.  My intent is to use those to safely port the C# over to JavaScript.

You can check the code out here.

If you prefer Java, Erlang or c, there is a repository from the original designers of the algorithm here.  A word of warning: if you try to use that repository to follow along with my code, it will be very difficult.  Conceptually, the code is somewhat similar to what I have written, but my implementation is almost entirely different.

Getting Started with RavenDB Using Pure JavaScript

RavenVsCouchYou might ask: “Why in the world would you create a pure JS app with RavenDB?”  I’m so glad we’re interested in the same things!  I’ve been toying around for a little while with CouchApp – which is a way to host applications completely within a CouchDB NoSQL database.  The idea is to greatly speed development and performance for certain application use cases by avoiding (most of) a server side middle layer.

Use Case

Let’s say you are building an application for internal use.  Assuming you are responsible on the client side, do you really need to have server side data validation?  I suppose you could find a few arguments in favor of it, but do they actually outweigh the cost of implementing a middle tier for this use case?  Really??

I’m going to pretend like you said, “No, Dave – by golly, you’re right.”

In looking at CouchApp, the first problem I ran into is that it’s hard.  Like, really hard.  I mean these guys are probably all into mod’ed out Linux distros and neckbeards and shit.  Which is cool, but the problem is that they are NOT into creating canonical, orderly, convention-based documentation/tests/examples that explain how the hell to do anything.  They are too “relaxed” I guess.  Instead you can do whatever you want, man.  For instance, they have all these different ways to do html rendering.  Half of them are outdated, and the other half are poorly demonstrated.  You get the feeling they are too smart and excited to let stuff mature for 2 months before moving on to the next shiny byte. 

(I’ll admit that last paragraph is probably unfair, but give CouchApp a few hours and see if you don’t feel the same way.)

The other problem I ran into is the sneaking suspicion that the whole thing is dead.  If you look at all the docs, posts and hubbub, it seems to center pretty tightly around 2010 and then tail off after that.  I tried to get some people from the community to give me some feedback about my last post, and all I heard were crickets.

The final straw for me on the whole CouchApp thing was that there is no easy way TO ACCESS THE DATABASE CROSS-DOMAIN.  Are you kidding me?  What the hell is the use of having a database that faces http if you can’t access the thing via http?  The solution is to install a proxy on your Apache server.  WHAT!?  I’m done.

Quoth the Raven

Enter the RavenDB.  If you compare www.RavenDB.net to www.CouchApp.org, it’s pretty glaringly obvious who has their shit together and who doesn’t.  I can hear my imaginary friend say, “Hey Dave, that’s not fair.  CouchApp is like, a side project, dude.  You should be comparing it to http://couchdb.apache.org/.”  And my friend would be right.  So go look at http://couchdb.apache.org/ then.  I guess it is better than CouchApp.org ….

And when that same friend then says –

“But Dave, RavenDB Costs Money and Shit”

– if he is truly concerned about RavenDB costing money, he should use CouchDB, MongoDB or Cassandra or some crap like that … (freeloader).  He should have fun with that.  I’m trying to get things done.  But hey … if my buddy really feels software should be free, then he should probably open source his own project.  Then he could use RavenDB for free.

Ok, maybe that’s a little harsh.  Maybe I’m being too hard on my outspoken pal.  But you know … the tone that I used was EXACTLY HOW I MEANT IT BE.

Brass Tacks

… As in it’s time to get down to them.  How the heck do you get going with RavenDB anyway?  Well the first thing to do is drive your browser over to Mr. Ayende’s shop and get yourself a build.

Ok, on another side note, does this guy Ayende or Oren or Auryn or whatever his name is kick ass or what?  I mean, I know he’s been putting out awesomeness for something like a decade now, but who decides one day that, “Hey, I think I’ll build a NoSQL database by myself.  Oh, and while I’m at it, I’ll make it the best one available on the market.  I’ll actually make it work well, have good documentation, be a (C#) developer’s dream to use, be easily distributable and you know what else?  If somebody sends a message to my mailing list, I’ll respond in less than 5 minutes even if I don’t know them AT ALL.”  Too bad he has a problem with run-on sentences.  Oh wait, that’s me.

Go over to http://ravendb.net/download and pick your poison.  Usually I prefer all things NuGet, but in this case, I didn’t want to find out whether or not the server is in there.  I just downloaded the zipped build.  Unfortunately, the most current build I found (960) had a bug with posting new documents using $.ajax.  This struck me as so egregious that I nearly didn’t write all those nice things above about Ayende, but stumbled in despair back to Couchappland.  Fortunately, the “unstable” build 2063 works … even if it does have some weird ass shit going on with a system database being the default and going completely paisley if you try to do the advanced database creation stuff …. It is labeled “unstable” after all, but I digress … again.

Once you’ve done the dance of unblocking the zip file and extracting it and all that, you can go to the Server subdirectory and type

raven.server /install

Congratulations.  You now have a running RavenDB service.  Beats the hell out of installing SQL Server doesn’t it?  You might also want to compare this process to that of CouchApp in my last post.

Oh wait, do I need to install a management studio?  No, it’s there already.  Just go to http://localhost:8080 if you don’t believe me.  Oh ok, do I need to install some configuration app?  No, you can just hack the config file.  And while we’re talking about it, why don’t we do some configuration file hacking?

Configuring RavenDB

You will find raven.server.exe.config in that same Server directory.  Open it with your favorite text editor and you will see something like this:

<?xml version=”1.0″ encoding=”utf-8″ ?>
<configuration>
  <appSettings>
    <add key=”Raven/Port” value=”8080″/>
    <add key=”Raven/DataDir” value=”~\Data”/>
    <add key=”Raven/AnonymousAccess” value=”Get”/>  </appSettings>
    <runtime>
        <loadFromRemoteSources enabled=”true”/>
        <assemblyBinding xmlns=”urn:schemas-microsoft-com:asm.v1″>
            <probing privatePath=”Analyzers”/>
        </assemblyBinding>
    </runtime>
</configuration>

 

Ok, I was actually a little surprised that port 8080 was available on my machine, so I changed that right away.  Also, I don’t want to fiddle with security right now.  Because I’m behind my firewall, I’m going to enable anon access on all interactions, and I’m going to leave Cross Domain Access wide open.  So now I have:

<?xml version=”1.0″ encoding=”utf-8″ ?>
<configuration>
  <appSettings>
    <add key=”Raven/Port” value=”49589″/>
    <add key=”Raven/DataDir” value=”~\Data”/>
    <add key=”Raven/AnonymousAccess” value=”All”/>
    <add key=”Raven/AccessControlAllowOrigin” value=”*” />
  </appSettings>
    <runtime>
        <loadFromRemoteSources enabled=”true”/>
        <assemblyBinding xmlns=”urn:schemas-microsoft-com:asm.v1″>
            <probing privatePath=”Analyzers”/>
        </assemblyBinding>
    </runtime>
</configuration>

 

Restart your service.  It’s in the Windows service.msc app or you can just type Raven.Server /restart from the command line in the Server directory.

Let’s Write Some JavaScript Already

Break open your favorite IDE/text editor, and because they all support NuGet, get yourself the QUnit-MVC package.  Or maybe they don’t, and you can get QUnit at www.qunitjs.com.  It’s hidden away down there at the bottom of the page for some stupid reason.

Now we need a test page.  Create an html file, and put this in it:

<!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”>
<head>
    <title>QUnit Test Page</title>
    <link rel=”stylesheet” href=”qunit.css” type=”text/css” />
    <script src=jquery-1.7.1.min.js” type=”text/javascript”> </script>
    <script type=”text/javascript” src=”qunit.js”></script>
         <!– App code goes here –>
    <script src=”app.js” type=”text/javascript”></script>
    <!– Unit test code goes here –> 
    <script src=”appTests.js” type=”text/javascript”></script>
</head>
<body>
    <h1 id=”qunit-header”>Intertwyne QUnit Test</h1>
    <h2 id=”qunit-banner”></h2>
    <h2 id=”qunit-userAgent”></h2>
    <ol id=”qunit-tests”></ol>
</body>
</html>

In a nutshell, this is what QUnit wants in order to display your test results  Obviously it might be better to put these scripts into special directories according to whatever conventions to which you subscribe.  If you actually did get qunit from NuGet, then you’ll need to square up your script and CSS URLs to match the Visual Studio conventions (duh).

Ok, now open up a new file called app.js.  In it, you’ll need to put something like this:

story = window.story || {};

story.url = “http://localhost:49589/docs”;

story.basicInsert = function (insertData, requestorCallback) {
    $.ajax({
        cache:false,
        type: ‘POST’,
        url: story.url,
        dataType: ‘json’,
        contentType: “application/json”,
        data: JSON.stringify( insertData),
        success: function (data) {
            requestorCallback(data);
        }
    });
};

story.basicGet = function (collectionAndKey, requestorCallback) {
    $.ajax({
        url: story.url + ‘/’ + collectionAndKey,
        dataType: ‘jsonp’,
        jsonp: ‘jsonp’,
        success: function (data, textStatus, jqxhr) {
            requestorCallback(data, textStatus);
        }
    });
};

These are a couple of JavaScript functions to write some JSON in and out of your RavenDB database.  I am using a POST rather than a PUT because I didn’t feel like finding a time sequential UUID generator for my IDs.  RavenDB will do that for me if I POST, sending back the results as JSON. 

The GET is requesting the results as JSONP so that my browser doesn’t freak out about cross-domain request results.  If you don’t know what that means, then Google it or ignore it because I took care of it for you.  This blog is already getting epic in length.

The other thing I do for both of these is pass in a callback parameter so our consuming functions can get the asynchronous results.  If you don’t know what a callback is, then consider a different vocation/hobby.

Ok, now onto the tests!!

Open yourself an appTests.js file and put something like this in it:

module(“TheRedCircuit’s tests for to show the good people”,  {
    setup: function () {
        // you can do some setup type stuff here
    }
});

test(“basicInsert testStory insertsIt”, 1, function () {
    stop(1000);
    var insertData = {name:”some title”,body:”some test body”};
    story.basicInsert(insertData, function (insertedData) {
        var key = insertedData.Key;
        story.basicGet(key, function (results, textStatus) {
            equal(textStatus, “success”);
            start();
        });
    });
});

Ok, so I’m cheating pretty badly here on the unit testing front.  I’m testing two functions at once, but seriously, how would you do it?  The chicken has to come before the omelet right?

When testing asynchronous functions, you have to tell QUnit to hold its horses while you go off across http land and do your thing.  That’s what the stop (and timeout after 1000 milliseconds) function is for.

Then I’ve nested all the calls so that we can only pass the equal function assertion if everything behaves nicely and gives us a “success” result.  Then the start function tells QUnit it can have the reins back.

Summary

I’ve done some whining about how hard CouchApp is.  I’ve verbally abused my imaginary friend.  Then I told you RavenDB is a lot easier because it is.  Then I showed you how brain-dead easy it is to get a RavenDB server going.  Lastly I did some POST and GET data access using jQuery.  Oh, and I showed you how to do some JavaScript unit testing with QUnit.

Because I know you are just falling all over yourself to know more, I’ll probably post a more complete version of this application next time, exploring RavenDB’s HTTP API some more … kind of like I did with CouchApp.

Mobility is a Mess

Lately, I’ve been trying to learn mobile development for both professional and open source coding.  The irony is that I’ve been doing mobile development for 8 years.  Unfortunately, mobile enterprise data acquisition has been tied closely to Windows CE/Windows Mobile for a long time.

Recently Microsoft created chaos in the market by abandoning support for Windows CE in Visual Studio 2010.  They also decided not to release any new version of .NET Compact Framework as a development platform.  When they realized how stupid this was, there appeared to be some sort of internal mess as MS tried to make up lost ground with the vendors that had sold bazillions of Windows CE units for them in the past.  After stumbling over themselves for a while, they’ve consolidated their embedded offering with their handheld offering to create Windows Embedded CE and pulled together some sort of backward compatible offering called Windows Embedded Handheld (at least that’s what they are calling it this week). At this point, I really don’t care very much.  I get the feeling that the market has been burned and will be going to Android.

Motorola Solutions is coming out with its first Android offerings now, and this means a lot of choices are ahead for the company I work for.  Should we attempt to write Java when we have never deployed a line of Java code in the past?  Should we use an offering like Mono for Android?   What if we want to offer applications in iOS as well?  Should we learn Objective C too?

Motorola Solutions is offering Rho Elements to allow developers to create out of the browser applications on their devices using JavaScript and HTML, but they are asking users to pay a steep per device license fee.  It seems like such an obvious losing strategy, I have no idea why they have even considered it.  Perhaps they aren’t aware of this widely used tool called PhoneGap that does (by all appearances) the exact same thing … for free.  Oh, and PhoneGap is affiliated with Adobe so PhoneGap gets first class support in DreamWeaver, and a high probability of continued adoption from the community.

PhoneGap seems like the obvious choice in this scenario, but there are a couple of other tools to consider.  Appcelerator is really starting to make a name in the mobile market by doing much the same thing as PhoneGap.  The difference between them appears to be that Appcelerator actually has a per-developer license fee associated with it, compiles to native code (making it speedier), leaks a lot of memory if not handled correctly and has relatively horrible documentation.  Also, it uses JavaScript only.  I haven’t figured out how the UI is created yet without HTML.  I don’t really feel like forcing our designer to use another IDE, or paying for the additional license for that matter.

The last option I’ve seriously considered is Mono for Android.  Any of the options in this article could probably fill an entire blog post with pros and cons, but Mono in particular is an interesting beast to me because I’ve come from a .NET background.  The main problem with Mono is that it is such a niche technology and is pretty much owned by Xamarin.  Although Xamarin has some very smart people running it, the company is less than a year old and I haven’t seen wide adoption of their technologies.  In the development blogs I read, there is rarely discussion of them at all.  It just feels like a big risk at this point.  The $800 per developer (to get both iOS and Android) is negligible relative to the potential development losses.

On the other hand, JavaScript and HTML, are free and being adopted by everybody.  In case you’ve been living under a rock for the past few years, JS isn’t a toy language anymore.  There are plenty of serious frameworks aimed at creating professional, maintainable code in JavaScript.  Windows 8 is favoring JS and HTML as the primary mode of delivering windows out-of-the-browser apps.  Adobe is killing off mobile Flash in favor of JS and HTML, Google uses little else, and Microsoft probably will not be offering a Silverlight version after (the current) v5.

How can you really go wrong?  If PhoneGap miraculously disappears, you can (obviously) run your app’s hardware agnostic code in the browser.  You can consider Rho Elements if you have to port your code as quickly as possible, or you can reuse your business logic in Appcelerator.  You can even reuse your client side business logic on the server if you like (using Node.JS).

Of course, the mobile development community changes so quickly that I could be laughing ruefully at this post next year.