scmple

Tagline - This should be easy. We think it should be simple.

Scmple is my (oops our) new pet-project. What I am referring to as this and it are common problems we all face working as part of a small team on a small project. We, as part of the scmple team, want to help you to iron out these problems. We hope to make it easier for the developers to spend more of their time and effort on developing software and less on (painfully time-consuming but much-needed) administration.

Here is a typical scenario that inspired us to come up with this idea - Person A has an awesome idea, it's going to be The-Next-Big-Thing! Person A codes away at night for a few days and brags about it to Friend B. B is now all fired up and volunteers his supreme comp sci development skills. So A and B work on this new side project (lets call it TNBT). Being developers, they know they have to use a versioning system and pick SVN or CVS or Git and they start coding away. They face a few hassles along the way - who works on what feature, who fixes which bug, is there a trail of all the bug fixes (aside from versioning history), who is tracking the project milestones, what is the future feature set..but they ignore them for the time being since having functional and working code is more important. Friend C also wants to join them. So as a small team, A, B, C have to collaborate on working on TNBT and to make progress, some of the hassles need to be addressed. Keep in mind that the only time they get to work on their dream project is after-work hours and on weekends. Hence the time that they actually spend on working together is very precious. They need to be able to use this time wisely and not in integrating and organizing three to-do lists with 100 bullet points and spreadsheets containing a complex matrix of which person is assigned to which task. What this team needs is a simple, straight-forward solution that offers them project management, task management, work assignments, issue tracking, discussion forums.

Scmple to the rescue. What we hope to achieve here is to simplify the process of product and project management. I know that developers dislike the word management, but to make your product even a tiny bit successful, you need to be able to manage it efficiently. Scmple will provide a hosted platform for versioning, with a built-in feature set like issue tracking, task management, project milestones, discussion and support forum. It's designed to integrate effortlessly with your development enviroment so that writing code and managing a project go hand-in-hand.

Technology - Open Source is the way to go. We decided to use Git as the version control system (read my blog post), Ruby on Rails and PostgreSQL, with a lot of AJAX goodies thrown in. We are looking at a variety of solutions and mayyyyyybe mootools is winning.

We are working on getting a beta version out soon. Will keep you guys updated!

Moving on to another job

This post comes a little late, but better late than never! After having spent a little more than a year at SITA, I decided to leave. Just yesterday, I finished my first 2 weeks at Premiere Global.

There were a number of reasons for this transition. I have had a good time at SITA, I like the people I worked with and I learnt a lot from them. The main reason for this move was that I felt like I wasn't challenged enough. I wanted to learn something new, work on something exciting. The projects that I had been working on for about a year were sort of coming to an end, and I guess I wanted a change. It was a little hard to leave since I am emotional and very loyal to the company but everybody understood my reasons to leave. So after a farewell lunch at a hibachi grill and a going away gift (a cute pineapple plant), my last day at SITA was over.

A couple of friends recommended Premiere Global Services (or PGI as we lovingly refer to it) to me. And hopefully vice-versa! I have been wanting to get into working in a web services environment and it sounded like PGI was working on re-designing their eMarketing to use a web-services-based framework. I interviewed with the team a few weeks ago and it seemed to be a good fit.

So here I am - 2 weeks into PGI and loving it! The guys in my team are real smart, I get to learn something new every single day and I have already written code and checked it in! Awesome!

Ubuntu vs Mac OS

So I recently bought an iMac, and I love it! I still have my dell laptop which has Ubuntu on it. Funny thing is that I recently switched over from OpenSuse to Ubuntu. IMO Ubuntu was easier to install from OpenSuse. Of course taking into account that I had installed OpenSuse on a purely Windoze box, all the fun stuff involving partitioning was already done for me. Anyway I did spend a few weeks with Ubuntu installing software and configuring my desktop so I do have some sort of an experience with it.

So If I have a laptop, why do I need another personal machine??
I have always wanted a machine with Mac OS on it. I used the Apples sparingly at Emory, but never truly got the hang of it. Ever since my husband got a mini, we sort of started replacing all our machines at home to either Apple TVs or Mac minis. I never really had the need for another machine since my Dell was good enough for all my side-project development machine. Or let me rephrase - I was happy with my laptop for the last 4 years. I do have a few issues with my Dell right now - I can't test out my web apps on Safari/IE/FF, Its not very fast (tops out at 2G RAM), I don't have a fancy text editor(tried to pimp up gedit to emulate textmate), and well I really wanted a change in my OS.

This last weekend was the tax free weekend. We stood in line at the apple store and were taken care of pretty quickly. After 20 minutes, I was the proud owner of an iMac **and** a free iPod touch! The iMac, designed by Apple in California, is a simply amazing design - the monitor has the entire computer inside it and its still not that thick. The speakers are not bad and I have enough USB ports on the back of the monitor - very sleek.

Once you get the hang of using an Apple machine with the funky Mighty Mouse, and learning your way around the Expose and the million keyboard shortcuts, its a pretty straight-forward OS to interract with. The only thing I might want to complain about is the lack of focus-follows-mouse support in the windows manager. Ever since my school days of playing around with Enlightment, I have been addicted to this very customized preference.

Anyway, other than that, here is a list of differences I see between Mac and Linux -

1. Ease of installation of software - On Linux, installing a software or application is not the most complicated process, but it takes time - the basic steps are downloading the binary or the tar ball, unzip it, make it, and then copy the files over to /the/proper/location. On an Apple, all you do is download the mac-os-version and copy it to Applications. Thats it - how easy is that.
2. The Dashboard - On Linux, you would probably need to start up all these applications in a different desktop - there is no dock like functionality. On an Apple, its always there at your beck and call . I like using it without the need to crank up the date app or the calendar app or the weather app.
3. Fluid - There is no way that I can explain how convenient SSBs are!! I have facebook, hahlo, jango, zimbra mail all opened up as its own little application. There are neat scripts that make the GUI look better and I don't have to bother about whether one app might cause a crash in another. So sure, I could open up different sessions of FF but it is not the same experience. Also, fluid is based on the Safari technology, its a lot faster.
4. TextMate - Eclipse is way too bloated to code. If I am not writing java code, Eclipse is more of a hassle to use than anything. Textmate is so easy to use, so many different templates and such a simple interface.
5. iTunes - I have tried to use amarok to sync my ipods. It hasn't been a complete disaster but its not intuitive. iTunes does it the best!
6. Prettier fonts/design/layout/shadows/GUI - Need I say more - its just prettier!
   

Hopefully I can code up the next-earth-shattering application on my new beautiful desktop, and then this purchase would be justified!

Grit Documentation

Grit is a Ruby library for extracting information from a git repository in an object oriented manner.

I am trying to use grit for a simple project - basically provide a GUI for git. Eclipse does a good job for CVS and SVN and gitk doesn't do a good job with git. It sounds like a good idea, and I really am all for git. Other than a few irritating things about merge, git is a pretty kewl tool.

So for this simple project, I am using mojombo-grit (and Ruby on Rails) to look at the tree/branch structure of git commits. Its pretty straight-forward but the API documentation is not that great. Its just not laid out intuitively and IMHO not easy to understand to follow. I like their example on the home page, which starts you off easily. But then after that, if you have to get more information on (say) a branch, then the doc doesn't do much for you.

Looking around for a better take on the documentation ....

GIT

A couple of reasons have led me to play around with GIT. I have had experience with CVS and Subversion. I use CVS at work, and use Subversion for my personal projects. Both have their pros and cons, and I heard a tons of good things about GIT. Linus wrote it and uses it for the kernel development - that's good enough!

After a ton of reading up on it, and watching
Randal Schwartz's presentation on git, I was inspired to get git and use it. It hit the spot - I have always had a few issues with the source code management systems out there and git supposedly has thought of everything -

• So here's a typical scenario - I checkout the source code from the HEAD at my work repository. I want to refactor it but don't want to break anything. I just want to change a few things, make them more efficient ( yea right! ) and basically try out a coupla things I couldn't try out earlier due to deadlines. So I edit code, play around with some kewl new technology, refactor it. By the end of the day, my editing has been finished. But its not perfect and frankly I rather not commit these changes since I really don't wanna break the code and/or I don't want the rest of the world to know I made these changes! So the only option left for me is to delete these changes ( or make a local backup copy ). I will have to start working on the main branch again, I guess I'll have to forget about this entire day's of work. The problem with that is that I have lost my work; however trivial it was. Hence the concept of a distributed version control system appeals to me. I clone the repo and play around all that I want. Ooo the concept of private branches...


• Another interesting feature of distributed development is that since you "clone" the repo to your hard disk, you always end up having the entire repo, the history of each and every file - all readily available. You can work on it offline, and commit it only when you are absolutely ready. This also helps when more than 1 person is working on the same piece of code, and needs to be able to make changes all the time.



• Every time I have to work on >1 different branches of the same project, it is always a hassle. I don't know of all the .cvs ignore files that are lingering around, and I always make mistakes committing my changes. I tried out git and tried out checking out 2 different branches. I tried out editing them, and then checking those changes in. Wow, that was simple! It really is very intuitive and the non-polluting of any sub-directories makes it so easy to learn and use git.

Git also delivers on a few other things - these weren't amongst my initial requirements, but are very useful now.

• For legacy humans ( ;) ), git has a "backwards-compatibility" feature to work with cvs clients. The end user does not need to learn 5 extra git commands, and can use cvs-style commits/checkout/branch to basically use git. I guess in my company, if I was trying to push Git, this would be an awesome plus-point!


• Git has funky merge and rebase features for tons of different users working on the same code. Merging is always a nightmare, and its not too bad with git. Of course you are going to break something somewhere, but thats what a developer in a team has to go through!


• It has to be real fast - the linux kernel has 25000 files

After using git for a few days now, I feel like I need to use it in one of my personal projects. I have been looking forward to working on something exciting.

Think Globally, Work Locally.....

OutOfMemoryError: Java heap space

Ugh memory leak! This is what I encountered a couple days ago at work. My java application is not too complicated - parsing data, normalizing it, and then inserting into the database. Its a kinda "critical" application and should always be chugging away. Its a a very random occurrence and I haven't been able to re-produce it yet. Sometimes I get this error after 1 day, sometimes 10 days, and sometimes none at all ( or like maybe at infinity ).

So one of the 2 causes of OOM is inadequate heap space - so I allocated more and ran the app, this time with the Xms and Xmx arguments. It ran for a longer time but again crashed with the OOM error. OK so time to look into the second reason - Memory Leak.

I looked around for a decent profiling tool, and downloaded JProfiler. It's free for 10 days and I hope I can figure out what the issue is before that. Installing it on my RedHat box was a piece of cake, but setting up the environment wasn't! Anyway, after setting it up, I started it. Way too cool!! ( Not the debugging/diagnostic part of the analysis ) I haven't really had experience with JProfiler so this is very new to me. The amount of data that is available is very impressive (and in pretty colors too ;) ). Tons of different views and graphs to help make the software more efficient and memory-leak free. Looks like I am going to be using the Heap stack view in my search for the memory leak.

Right now, I am familiarizing myself with this tool, reading up on what should be my best approach. It's more of an art to find the runway non-GC'ed variables than science! A couple of things I am going to take a look at - objects like String, HashMapEntry, Char, Long maybe a problem due to their growing numbers (but maybe thats valid ?), database connection pooling and result sets not getting closed properly....

I am looking forward to playing around with JProfiler, but not as much to finding this "random" memory leak.