If you wonder who Joel is, then you're faulting again because this mean that you're not reading his blog, and let me tell you, you should really invest some work time to read it. Joel Spolsky is an experienced programmer, manager and business runner, and above everything, a kind enough person to share his knowledge online. Days after days, his blog became one of the most read and relevant source of inspiration for software developers. So much, that the blog was finally compiled up to edit the first edition of "Joel on Software", and now the follow up, "More Joel on Software".

So this second book is made of nine sections that each address a concern of the software world:

• Managing People: If you ever had the opportunity to manage software developers, you know how hard it is. Joel goes through management and recruitment techniques, why they don't work and how you should proceed.
• Advice to Potential Programmers: Programming beginner / student ? Read these lines carefully, Joel is right on the whole line: Attend hard courses and get over them, this is mandatory.
• The Impact of Design: Sharp and relevant guidelines on several segments of software design, from style to features.
• Managing Large Projects: How Microsoft failed at managing big projects over time, why it's hard and how you should deal to still be happy at the end of the day
• Programming Advice: One of my favorite sections. Joel unveils his secret to achieve efficient software scheduling, the Evidence Based Scheduling method, how to make wrong code look wrong etc...
• Starting a Software Business: Two relevant and almost moving forewords written by Joel for two other books, telling you why and how you should start a software company, what you should focus on. If you have this little entrepreneur seed inside of you, you'll love each line of this section.
• Running a Software Business: This is the hard part of the book. Joel describes how they designed their offices to be the best place possible for software developers, making you feel like you're working in a rat cave, even if you work in front of one of Paris' most beautiful place
• Releasing Software: Once again, invaluable advices on how to pick a release date and decide on your pricing method. Very accurate.
• Revising Software: My previous boss should have read that part. How NOT to refactor your code, efficiently solve outages and avoid them to append again with the five whys, and to end the book, a method to establish relevant priorities time-lines.

So yeah, this book rocks hard and is, to me, a must read. Each rant and each advice feels (and hopefully is) weighted by a strong experience, and given to you so you don't have to wait years to learn this on your very own. Finally the best thing about "More Joel on Software" is that it's written as a blog, in a lightweight, (VERY) funny and sincere way, not as an overworked speech.

Oh and no, I'm not sponsored and I won't get anything if you buy the book. If the book was a turd, I'd totally tell you too

Jul 12, 2009 6:26:29 AM com.google.apphosting.utils.jetty.JettyLogger info
INFO: jetty-6.1.x
Jul 12, 2009 6:26:29 AM com.google.apphosting.utils.jetty.JettyLogger warn
WARNING: failed com.google.apphosting.utils.jetty.DevAppEngineWebAppContext@c45809{/,/Users/work/workspace/Athena/war}
javax.xml.parsers.FactoryConfigurationError: Provider org.apache.xerces.jaxp.SAXParserFactoryImpl not found
    at javax.xml.parsers.SAXParserFactory.newInstance(SAXParserFactory.java:113)
    at org.mortbay.xml.XmlParser.(XmlParser.java:69)
    at org.mortbay.jetty.webapp.WebXmlConfiguration.webXmlParser(WebXmlConfiguration.java:83)
    at org.mortbay.jetty.webapp.WebXmlConfiguration.(WebXmlConfiguration.java:78)
    at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
    at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:39)
    at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:27)
    at java.lang.reflect.Constructor.newInstance(Constructor.java:494)
    at java.lang.Class.newInstance0(Class.java:350)
    at java.lang.Class.newInstance(Class.java:303)

And guess what ? This is actually a Java (jre) 1.5 bug ! So the easy solution is to switch to Java 1.6 which also allows you do to String.isEmpty() instead of String.length() == 0 which is just awesome (...). So you switch to Java 1.6, and try to start your app again, and then cry again when you get this clear one-line message that haunts you since you've read it:

You must use a Java 1.5 runtime to use GWT Hosted Mode on Mac OS X.

Damn.

How the hell can you get this stupid app running ? The solution is damn easy but again there's a trap. One thread on stackoverflow tells you that you should replace you gwt-gears.jar file with gwt-gears-noredist.jar that is available in the GWT Gears API archive, and tada. So you go on the GWT Gears API website and get the gwt-gears-1.2.1.zip file in the download box on the right side, and of course, there's no gwt-gears-noredist.jar file inside the archive. This is the moment where you get a little angry and depressed. Where the heck to find this file ? Well, don't ask me why, but the gwt-gears-noredist.jar file is only available in the tar archive, known as gwt-gears-1.2.1.tar.gz, that you can only find in the Downloads page of the site.

This is fucking stupid and make you feel like you've lost precious time because, so I hope this post saved you that

• OCZ to sell the Z-Drive p84 PCI-Express SSD, a high performance flash drive on top of a PCI-Express architecture with impressive specifications: Read up to 870 mb/s and write up to 760 mb/s (traditional hard drive are around 60 mb/s for both write and read). The p84 will be available in three capacities: 250Gb, 500Gb and 1To (from $1,561$ to $3,369). (via Mac
• 67 Terabytes for $7,867: Petabytes on a budget. This amazing article has been retweeted all around the web and deserves it. Backblaze provides simple and accessible online backup with unlimited storage for only $5 / month. To do so, they had to figure out how to store hundred of petabytes in a reliable, scalable and low cost way. They describe in details how they built their own solution, really cool. (via twitter)
• Kernel 2.6.31 to speed up Linux Desktop: Cool announcement as I was really getting upset with my work station on Linux: "As the Linux community looks forward to another kernel release, the kernel hackers have been working on improving the memory management so that the X desktop responsiveness is doubled under high memory pressure. The result is an improved desktop experience. Benchmarks on memory-tight desktops show clock time and major faults reduced by 50 per cent, and pswpin numbers (memory reads from disk) are reduced to about one-third." (via Slashdot).
• Software Pricing, Are we doing it wrong ? A quite old post (I'm one month late on this one.. :/) but definitely worth the read. A simple quote to tease you: "the idea that software should be priced low enough to pass the average user's "why not" threshold is a powerful one."
• Other interesting things: TinyEye, search images by uploading one (doesn't work that well unfortunately..). Yet another Hadoop post: Hadoop continues to excite the cloud.

The average quality of the selected papers was pretty good, but I personally thought that there were still too many non innovative papers and badly prepared speakers. Common guys, you've got the chance to expose your work to an invaluable audience in a prestigious conference, the least you can do is to be well trained so you don't have to skip your last 10 slides... Also, another detail, how the hell can you receive 700 persons in France and serve lunch without desert ? I may be a bit greedy, but I wasn't the only disappointed person. That's it for the traditional french whine part. (But there was good wine !)

So the last two days were full of interesting sessions in various domains. Wednesday morning began with a keynote about how database technologies contribute to enhance games and simulation engines performances. For a keynote, I would have preferred a more database focused talk, because in the end it was just about having each tick of the game acting in a map / reduce fashion to compute all changes together, and some Tree usage here and there, nothing very funky (but there were some great videos of old video games to compensate... ).

The afternoon started with a great session called Map/Reduce with three presentations:

• SQL/MapReduce: A practical approach to self-describing, polymorphic, and parallelizable user-defined functions: The speech was about a commercial product, AsterDB, and their alternative approach to Hadoop / PIG. The idea is that statisticians and other people interested about processing data prefer to write SQL queries instead of coding map / reduce functions that are usually so specialized that they are not reusable. They answer this concern by providing an easily extensible SQL language so you can bypass the actual limits of SQL by writing a few lines of java for example, the whole thing being then parallelized.
• Building a HighLevel Dataflow System on top of MapReduce: The Pig Experience: Really interesting presentation by a Yahoo! engineer, one of the creator of Pig, the famous high level language to express data analysis programs on top of Hadoop. The focus was on how Pig is designed and implemented to actually transform your logical instructions into physically distributed jobs and stages. Pretty neat.
• PLANET: Massively Parallel Learning of Tree Ensembles with MapReduce: Highly technical talk from a Google fellow describing "a scalable distributed framework for learning tree models over large datasets. PLANET defines tree learning as a series of distributed computations, and implements each one using the MapReduce model of distributed computation.". A great breakthrough in a domain where lots of state of the art learning algorithms are designed for a single machine.

The day ended on a pleasant discussion panel about "How Best to build Web Scale Data Managers ?". Lots of relevant questions and responses that point to the disappointing lack for an open source parallel database. Hopefully the next decade will solve that. (edit: I might not have understood some of the subtle sarcasm of one of the speaker, so lets remove the critic).

Thursday was even more interesting, starting with the "10-year award keynote", rewarding the most influencing paper of the last decade. The award was attributed to the MonetDB staff for their paper "Database Architecture Optimized for the New Bottleneck: Memory Access" describing a novel approach to database storage (column storage, well known nowadays) and to the join operation optimized for the CPU cache, thus avoiding the overhead of memory access, and how the column oriented engine allow for . They also talked about vectorwise and its enhanced query computation pipeline performing batch hit operations. Well done guys, it's a pretty impressive work !

After a little coffee i attended a really great talk given by a postdoc from ETH Zurich (i don't know how many of these guys came to the conference, but there were a lot !) about "Data Processing on FPGAs". I didn't knew much about these reprogrammable chips before, but the guy made them look pretty cool. The presentation described how you can achieve faster and parallelized sorts and actually save money because of the incredibly low power consumption of these chips (8W vs 102W for a CPU). (And the guy was well prepared and gave a dynamic and fun presentation !).

Another trendy topic followed: HadoopDB (yeah, there where Hadoop and map reduce all around the place). HadoopDB is "an architectural hybrid of MapReduce and DBMS technologies for analytical workloads", a work done by a few students of the Yale university. Their approach is to reconcile the two elephants together (Postgres and Hadoop both have elephants as logos), put an modified version of Hive on top of that (Hive is a SQL => Map Reduce job interface) and make use of the processing force of the databases to actually do what they are meant to do in a distributed fashion. It's an interesting track even if it feels a bit like a big hack on top of hacks. And erm, i can't say much of the end of the presentation, we havn't seen the last 10 slides or so... :/

Finally, I ended my visit at VLDB09 with two presentation of Google Interns about data mining to get structured result sets out of semi unstructured pages with lists and tables. The papers are really neat and worth having a look:

• Harvesting Relational Tables from Lists on the Web: A method to detect fields in unstructured lists and to successfully align them up into a coherent table.
• Data Integration for the Relational Web: A description of a search engine that creates clusters of related tables on the web, rank them and is able to create a single coherent table with the similar fields, and to actually extends the table with new fields on demand that may not have been directly included in the original tables. A really interesting approach that sounds like google squared

Allright, I think i'm done with this dammnnnn long post that you probably don't have read till the end, I don't care it's also a way for me to keep track of all these interesting stuff

VLDB guys, congratulations and see you next year at Singapore !

This week I am attending Very Large Data Base 2009, a major meeting for people that deal with huge amount of data or concerned about databases performances. This event takes place each year in different cities, but we have the chance to host the 09 edition in Lyon, gathering all these brilliant people in France. Almost 700 attendees from 44 countries came up here and so far it was worth it.

The conference takes place till Friday, so I'll only share my thoughts on the past two days. First, the event is pretty well packed up. Everything from the place, the food, the quality of speakers to the variety of selected papers is pretty attractive. As usual, you are given a guide of the conference, but this one is particularly well made, with every talks abstracts, map of the city, side activities, etc... Special cool thing: All the papers are available on a USB key available in your attendee package, thus saving trees, time and money. Pretty cool.

Yesterday was a day of workshops. I attended two of them, the USETIM (Using Search Engine Technologies for Information Management), and the BIRTE (Enable Real-Time Business Intelligence). USETIM was chaired by Gregory GREFENSTETTE, Exalead's Chief Officer of Science. We saw some very interesting presentations, including big actors like Microsoft, with the demo of their Symphony platform that allows building a custom search engine (both in terms of sources and display) in a simple drag and drop interface. Unfortunately, the two talks of BIRTE I attended were not that interesting, mostly driven by industrial products.

Today began with a keynote of Raghu Ramakrishnan, Chief Scientist for Audience and Cloud Computing at Yahoo!, that talked about key-value stores, one of my favorite subject. He gave a presentation that could have been the full length version of my talk at Ignite Velocity 2009 (San Jose): Why these simple stores were created, what are the trade off of using such technologies, the details about Yahoo! PNUTS implementation and a comparison with other systems.

Another highlights of the day include a presentation from another Yahoo! fellow about "Indexing Boolean Expressions", a session on Data Visualization (largely inspired of the excellent Visual Display of Quantitative Information), and finally a paper describing a good alternative to B+Trees (a well known data structure in indexing technologies) for Flash devices and SSD drives: Lazy Adaptive Trees. Congratulation to these authors.

Beside that I really like the city, Lyon is a beautiful place ! The borders of the Rhone river are lovely and full of joyful people enjoying a little glass of wine after work, very pleasant

Now that I'm almost done with college (a few months left !), I can proudly say that this school brought me one of the best practical learning available nowadays, but unfortunately it miserably lacks of theory and algorithms courses, so I'm back reading books to try to fill a bit of this missing knowledge by myself (so mainly geek books and thrillers ). ANYWAY ! All that life stuff to prepare you to see some book reviews from time to time

I'll start with "Even Faster Web Sites", Steve Souders latest book. If you're a bit involved into web development, you probably know who Steve is. He's the guru and best evangelist of front end performance problems and solutions, and a pretty nice person working for the big G. He's responsible for at least 30% of savings on the load time of all the major web sites you're using everyday.

I totally loved his first book, "High Performance Web Sites", a set of 14 rules to speed up your web site on the client side, something that was quite neglected before his insights. The book was very concise, with succinct examples and no extra fat.

His second book, also knows under the little name of "EFWS", follow the same format but with 6 chapters contributed from some front-end engineering rock stars. It's a pretty good initiative as it bring directly to the reader the loss less speech of experts on each domain. Steve also decided to write a blog post for each of these rules on his blog so everybody can benefit of his knowledge and hard work without having to buy the book, which shows his devotion on making the web a better place.

So here comes the time of my raw (and a bit rude) thoughts on "EFWS". Overall, the book is very good and brings another set of important rules that each web developer should know about (the chapter about CSS performance is surprising!). Unfortunately, unlike his first success, some chapters of "EFWS" really feel repetitive, as he had to fill the pages to get the book done. I would say that 35% of the book got me a little bored, but the 65 other percents were just all right, particularly the ones on Javascript internals, flushing early and CSS.

If you're a decent web developer and haven't read the first book, HPWS, well you're not a decent web developer and go grab a copy RIGHT NOW. Otherwise, I definitely recommend going through the blog posts on Steve's website, worth the read !

• PunyPNG: A fresh view on image optimization that will hopefully save you the hassle of using separate tools. "punypng is serious about image compression — it handles 8-bit PNGs, 24-bit PNGs, JPEGs, GIFs and animated GIFs. It also leverages dirty transparency techniques to further optimize transparent images beyond what typical compressors like pngcrush can handle." (from Ajaxian.com)
• Google Cafeine: What is really is: A still high level but interesting post on what may be the underlying technologies and motivations behind the brand new Google Cafeine. (Obviously GFS2, real time web etc...)
• SQL Pie Charts: Probably one of the most useless thing on the world, but this is so awesome that the guy deserves a day of glory. Doing ASCII Pie Charts out of pure SQL is just insane.
• TomTom for IPhone released ! We're finally getting a real suitable GPS software for the IPhone. Doesn't seems to be available everywhere yet, but should be soon (from Mashable)

Some of you won't even believe this, but yeah, after more than a year without blogging, i feel like writing and sharing again. This blog will be much more technical than the defunct unixaumonde.com, so some of you may just get lost reading my thoughts, sorry

As your might have guessed I'll mainly talk about my passions, so a lot of search engine related stuff, information retrieval, natural language processing, distributed / cloud computing and web development.

Keep warming up folks, more to come soon !