Breaking news: TechCrunch is announcing that Microsoft withdraws Yahoo bid

TechCrunch is publishing a breaking news on the evolution of the Microsoft-Yahoo story that has been going on for the last couple of months.

I will talk here about the details of what it looks like the end of the Microsoft-Yahoo saga, but rather jump to the second part of the TechCrunch post which is presenting a letter from Steve Balmer (Microsoft CEO) addressed to Jerry Yang (Yahoo CEO).

There are a couple of very interesting aspects in this letter:

• Steve Balmer starts by thanking everybody from Yahoo that has been part of the discussion, but shortly afterwards he moves to describe the financial details of the disagreement
• later on the letter, Steve presents in 5 clear points why a deal between Yahoo and Google would make the acquisition "undesirable".

So, even if the story seems to come to an end now, I am getting the feeling that this is not the end and we may be seeing some more episodes in the future

Advertisement

CouchDB moving to Java… no, but FeatherDB is here

I am not sure if you got the April’s 1st CouchDB announcement:

So, after a lot of heated discussion internally, I’m somewhat sad to say that Erlang will not be used in future releases of CouchDB. We are switching the whole codebase to (…drumroll..) Java.

I don’t know how many have fallen for it (must confess that even if I am an easy target this one didn’t work for me), but reading something like this on April 1st is not always trustworthy.

However, a couple of days later a part of the announcement is becoming reality: FeatherDB – Java JSON Document database

According to the post:

FeatherDB was my attempt at a Java clone of CouchDB.

but

There were a few things about the way CouchDB handles documents that I didn’t like, so FeatherDB does things a little bit differently.

You should check the original post as it contains a lot of more details about what and how FeatherDB does and how it is comparing with CouchDB

Posted by: Alex Popescu (aka the_mindstorm) (see twitbits)

Extensive list of Amazon services resources

Tools

• Simple Command-Line Access to Amazon EC2 and Amazon S3: + project page
• EC2
  • EC2 AMI Tools
  • EC2 Command-Line Tools
  • EC2 AMIs
  • Cleaning up EC2 Images, AMIs, and Instance
  • EC2 New Instance Types
• S3
  • Amazon S3 Firefox Organizer S3Fox
  • Bucket Explorer
  • S3tools project
  • Jungle Disk
  • S3Drive Windows app
  • S3Browse
  • CLI to Amazon S3 Service using Ruby
• A MySQL Storage Engine for AWS S3

Backups on S3

• Why I Started Using Amazon S3
• Backing Up Flickr Photos with Amazon S3 (Python scripts)
• Amazon S3 PHP Class Update (PHP)
• File Uploads to Amazon S3 the AJAX way thanks to Dojo
• How I automated my backups to Amazon S3 using s3sync
• S3 Backup – Fail-safe, Encrypted Online Backup

Amazon S3 Resources

• Amazon S3: 16 Sites and Resources, Example
• CacheFly vs. Amazon S3
• Using Amazon S3 as in Image Hosting Service
• Introduction to Amazon S3 with Java and REST
• Automated Server Pool Management in Java
• Running Hadoop MapReduce on EC2 and S3 + Sample code

Amazon SimpleDB Resources

• What You Need to Know about Amazon SimpleDB
• Amazon SimpleDB – Limited Beta
• Amazon SimpleDB Developer Guide
• Amazon SimpleDB. What nobody is talking about?
• Eventual consistency
• Relating to SimpleDB
• Query 101: Building SimpleDB Queries
• Query 201: Tips & Tricks for SimpleDB Query

---

Amazon S3 Tips & Tricks

CNAME + domain name

You can create a CNAME record off your domain and point it to s3.amazonaws.com, creating a branded URL to your S3 bucket. Your bucket name must also match the CNAME + domain name.
So, you could create a CNAME record named “S3” and point it to s3.amazonaws.com. Then, create a bucket called “s3.codinghorror.com” in your S3 account and place all your static stuff there. Then, when you create the reference in your code you can use “s3.codinghorror.com”.

Virtual Hosting of Buckets

Virtual Hosting of Buckets

Past issues

S3 was designed for occasionally sharing large files, not “often sharing” small files. Ping times can be over a second, sometimes 2. It is great for 80 meg files every once in a while, but not for 15k PINGs 100,000 times a day.

S3 still seems to have some problems with speed. When serving greater than 60,000 hits per day, the images load VERY slowly. Except on days where it’s so slow it’s not even loading. We like the idea of s3, but we switched back to hosting our own images on likebetter.com after too many speed issues. Also, about 1 in 100 loads fail randomly.

---

Amazon Services Code

Java

• General
  • typica – An Interface to Amazon SQS, EC2 and SimpleDB in Java + typica
• EC2
  • EC2 Library
• S3
  • JetS3t (Amazon S3 library)
  • S3 SOAP Library
  • S3 REST Library
  • S3 Eclipse plugin
  • S3OSCache to Store Cached Data on S3 + http://www.lokitech.com/projects/s3oscache/index
  • S3 Key Management
  • S3Shell in Java
• SimpleDB
  • ‘Simple’ SimpleDB code in single Java file/class (240 lines)
  • Java Library for Amazon SimpleDB
  • JPA implementation for SimpleDB + simplejpa
• SQS
  • Amazon SQS Sample in Java
  • REST SQS Demo

Python

• boto: AWS for Python + boto
• EC2 Query
• S3 REST Library
• Monster Muck Mashup – Mass Video Conversion Using AWS

Ruby

• General
  • RightAWS + project page
  • RightScale Ruby library to access Amazon EC2, S3, SQS, and SDB
  • http://ruby-aws.rubyforge.org/ruby-aws/
• EC2
  • Amazon EC2 Ruby Gem + http://rubyforge.org/frs/?group_id=2753 + http://amazon-ec2.rubyforge.org/
  • EC2 Query
• S3
  • Amazon S3 Library for REST in Ruby
  • s3sync.rb and s3cmd.rb tools
  • Using S3 from EC2 with Ruby
  • AWS::S3 S3 REST API
• SimpleDB
  • Using SimpleDB and Rails in No Time with ActiveResource
  • Using SimpleDB with Ruby on Rails (ActiveResource)
• SQS
  • SQS – A Ruby interface to Simple Queue Service + http://sqs.rubyforge.org/
  • Amazon SQS Sample in Ruby
  • Basic Amazon SQS Sample in Ruby
    

Challenge: which Javascript library is the best

Lately, I’ve been looking again at the Javascript generic libraries market. And it looks like instead of stabilizing, more and more solutions are available. I do like having options, but at the time you have to pick one things are becoming much more complex. This remembers me of the status of web frameworks in the Java land, and I guess everyone agrees that things are quite messy there.

So, what criteria are you using when picking the solution to be used in your project? Is it your existing knowledge? Is it the quality of the documentation? Is it the maturity of the project?

I have gathered together a list of major libraries. They are listed here in alphabetic order for the moment, but if you start providing pros and cons about them the list will evolve to reflect your feedback. So keep the comments pouring!

1. jQuery (+9/-0)
2. Ext JS (+5/-2)
3. mootools (+3/-0)
4. mochikit (+2/-0)
5. prototype (+2/-0)
6. dojo (+1/-0)
7. qooXdoo (+0/-0)
8. Rico (+0/-0)

I will start by saying that I am aware of the quantity of (electronic) ink that have been used covering this subject on the blogosphere so far, but if you help me out with comments/details/opinions (subjective opinions are welcome too 😉 ), I am willing to put together a list of pros and cons for each of the above listed libraries.

Posted by: Alex Popescu (aka the_mindstorm)

Update 2008-03-20 12:46: You may wonder why ExtJS with +5/-2 is currently ranked before the mootools with +3/-0. Thing is that one of the -1s is about their licensing model, and as nobody here is a lawyer, we are currently awaiting for more details.

Posted by: Alex Popescu (aka the_mindstorm)

Mixins and Traits (for Java)

I’ve read last night a post on traits for Java and I just thought to make a small comment.

According to the author:

The difference between a Trait and a Mixin is that order is important.

Well, in my opinion there is a much bigger, fundemental difference: mixins have state, traits don’t. When a class or object includes a mixin it includes all methods and attributes defined by the mixin, according to the Wikipedia:

A mixin can also be viewed as an interface with implemented methods. When a class includes a mixin, the class implements the interface and includes, not inherits, all the mixin’s attributes and methods.

This is quite an important difference, as traits are easier to be supported (except the inheritance change they are imposing), but the benefit is really limited. And the good news is that using AOP (e.g. AspectJ) you can have mixins (so implicitly you can have traits).

Posted by: Alex Popescu (aka the_mindstorm)

JXInsight: how to document a rich UI app?

One of the suggestions I’ve had for improving JXInsight documentation was to create a small document describing the application (called JXInsight console) UI. Basically, this document should contain some short descriptions for the panels and also a legend for the icons used. In my experience, this approach proved to be the simplest, the cheapest but most effective.

But, it looks like this will not work (or it will just partially work) for JXInsight. The console UI (doesn’t this sound weird ? 🙂 ) is using around 1000 icons (yes, you heard me right). So, creating this document is gonna take a lot of time and is gonna be quite expensive to maintain. I am wondering what other options would you suggest for documenting a very rich UI?

Here is a screenshot of the app, just to get an idea of what I am talking about:

Posted by: Alex Popescu (aka the_mindstorm)

JXInsight documentation

After the other day William’s comments, I decided to invest some more time in reading about JXInsight. So last night, I’ve started my look through their blogged documentation and put together everything that might get me started. Thinking that others may be facing the same issue as I was, I am listing here those entries that may get you some ideas about the product.

101

• Services (description of concepts)
• AspectJ(Usage and extensibility of JXInsight based on Aspectj)
• Introducing Diagnostics

Configuration

• JXInsight and Tomcat 6
• JXInsight and WebLogic screencast

I have listed only the link for Tomcat as that is the configuration I was looking for, but it looks like the blog contains entries for other deployments too. There are a couple of more links that may be interesting, but they look a bit old: Apache Tomcat Monitoring and Configuring Servlet Traces.

My continued quest for information resulted in finding a couple of more pages, but I’ll have to filter them out firstly:

Concepts at work

• JDBInsight
• Probes
• JVMInsight
• Diagnostics
• Metrics

What I am still looking for, and doesn’t look like being around, are the following pieces of documentation:

• Definition of the concepts
• Console details
• Terminal usage

I’ll probably ping William about these. So far I’ve been a bit puzzled by the terms the documentation is using. These are not clearly defined, and I think I am that kind of person that agrees with usability guru Steve Krug when thinking that you shouldn’t introduce new terms for existing concepts, as these are usually leading to confusion.

Last, but not least, I have noticed that JXInsight 5.5 was released and that William will focus on creating the so-needed documentation:

Now the hard part begins – rebuilding the documentation set! This will be ongoing for the next few weeks[…]

I am looking forward to seeing it.

Posted by: Alex Popescu (aka the_mindstorm)

Distributed VCS: Mercurial

Lately, I’ve found myself using locally more and more Mercurial (a fast, lightweight Source Control Management system designed for efficient handling of very large distributed project). The reason is that I’ve been traveling a lot and meanwhile I had to play with some crazy ideas I had. And so far, I am very happy with it; even if I am missing a couple of things that would make it absolutely lovely (a decent Eclipse plugin, a GUI tool, etc.). So, if you know about such tools please let me know.

Yesterday, I have imported locally a CVS repository. Before doing it, I had to search a bit on how to do it, and below are the links that I’ve found interesting. I have ended by using the generic Tailor as I’ve used it before and it looked like the simplest setup.

Links

• fromcvs: fast (incremental) cvs->* conversion
• convert-repo
• Tailor (configuration details here)
• Tools for moving from CVS to Mercurial

I had also to install cvs on my Mac machine. It comes inside the XCode app, but I didn’t want to have all those, so I have found a repository of cvs distros.

Configuration

If somebody wonders what configuration I have used with Tailor, here it is:

[projectname]
root-directory=~/workspace/
source=cvs:projectname
target=hg:projectname
subdir=projectname
verbose= True

[cvs:projectname]
start-revision = HEAD
repository=:pserver:anonymous@projectname.cvs.sourceforge.net:/cvsroot/projectname
module=projectname

[hg:projectname]

Basically, this configuration says: import the project called projectname from the CVS source detailed in configuration section [cvs:projectname] into a Mercurial targer (hg prefix) detailed in the configuration section [hg:projectname].

Concerns

Even if I am getting more and more familiar with Mercurial, there is one thing that concerns me a bit: as long as there is no remote replica of my repository, I still have to deal myself backing up the system. Mercurial repositories are self contained, so if you by mistake delete the workspace then all your work is gone. Forever. I am trying to figure out if there is an easy solution for this.

Posted by: Alex Popescu (aka the_mindstorm)

Glassbox and JXInsight

I have heard about JXInsight a couple of months ago, and since then I was eagerly waiting for the moment I’ll get to it. Last night, with a bit of energy left after a long day, I have remembered about it and I thought: “maybe this is the right moment”. I have quickly passed through the homepage (which definitely looks interesting in terms of features) looking for a download. I have found the 5.5-RC1 and got it installed on my machine. When all this was done, I have started to look around to what’s next. And unfortunately I couldn’t find any good documentation about what am I supposed to do. The site doc links to a blog, where, even if there are various entries, none is close to being a decent user guide/step-by-step/up to speed intro. I have double checked the installation, and there was no documentation their either. Maybe I should use the examples… I have to confess that unfortunately, I had to give up as my energy was finished before figuring out what’s next. In conclusion, I’ve been quite disappointed that after a very cool homepage intro there was nothing to get me started (even if I have spent a couple of hours). Or maybe I was just missing the obvious?

Speaking about JXInsight features, I have remembered about Glassbox: an open source project created by my friend Ron Bodkin. In fact at some moment in time, I have spent quite a few weeks understanding the internals of the app and figuring out and suggesting some improvements. Now, Glassbox is in 2.0 RC1, and I hope the guys have addressed all the problems, because the product looks really interesting. And you know what? It comes with a decent User guide! Not to mention that it is open source, so at any moment, you can just stick your nose into code and figure out how is working.

I am wondering if others have any experience with these tools and if they are up to recommending something.

Update I have started to gather JXInsight documentation.

Posted by: Alex Popescu (aka the_mindstorm)

CSSVista… what’s that?

Well, it is definitely not a stylesheet editor for your Vista Windows OS. In fact it has nothing to do with Vista OS, but only with CSS.

It is a free product that should provide you the ability to edit live the CSS of your site on both IE and Firefox. That’s sounds cool! Editing live sites on Firefox has been available for a long time through the usage of Web Developer extension or the great Firebug plugin. But nothing similar, not even close, was available for IE. Now having this product will hopefully make our web dev life a bit easier, and having it for free is even great. Congrats to the Litmus guys!

You can find the product here.