Review of three books for web application development, by Lee McK
08/21/07

RailsSpace Building a Social Networking Web Site with Ruby on Rails, by Michael Hartl and Aurelius Prochazka, published by Addison-Wesley Pearson Education, c. 2008. $44.99 US.

I have already invested about 10 hours reading and working with this Ruby on Rails web site construction guide. Last month I reviewed the very similar Apress book Practical Rails Social Networking Sites.

See review at: http://slashdot.org/~beachdog/journal/177633

Both "RailSpace" and "Practical Rails Social Networking Sites" are second books for a person who has worked through one or two introductory Ruby on Rails books. The authors of RailsSpace point back to Agile Web Development with Rails by Dave Thomas and others as the most widely recognized first Rails book.

The authors of RailsSpace tell the story of how both of them have gone through about 10 years of web site development projects. They were both college graduate students that came upon an available web server. They moved from static html pages, then they embedded calls to Perl scripts in their html, then they worked with Philip Greenspun's Arsdigita Oracle database scripted in TCL running on the Aolserver web server, then they worked with the Python based Zope website system. That trail of experience leads to Ruby on Rails.

One of the things I like about RailsSpace is the authors take the time to very carefully focus on some of the conventions and features of the Rails environment. One of the first things you do with Rails is run "script/generate controller ...".

This book takes the time to point out the important things this action does. This script creates a file structure, some skeleton programs, skeleton display pages and writes some Ruby code snippets.

So RailsSpace spends more time at the beginning explaining the environment that Rails creates. More than being just a powerful script, the authors begin showing how Rails has constructively moved past the agony, the verbose code, and the blizzard of ticky tacky details that are common in the older development environments.

I liked the explanation approach because it has helped me to push on to developing my own idea of a social networking web site. I don't want to be a script copier, I want to prototype and experiment with a different design in parallel with working through the book.

Something I have found missing from the cook book type Rails books is assistance with the rhetoric, psychology and philosophy of designing a social networking web site. I like the formulation from Donald MacKay's Information, Mechanism and Meaning book; information is that which changes people's readiness for action.

The design problem of making a website that prepares a person for the action of buying something is thoroughly explored by many ecommerce web sites. The design problem of preparing a person for action to improve the quality of preschool childcare is not well developed.

Two books for the design problem:

Deliver First Class Web Sites 101 Essential Checklists, by Shirley Kaiser, published by SitePoint Pty. Ltd., c. 2006. $39.95 US.

This is a guide to designing and delivering a web site project. This book bridges the gap between a customer who wants to hire the construction of a web site and the web site creator.

The first two chapters frame the extremely difficult customer side problem of who is the audience, what is the purpose of the site and what is a realistic budget. The way the book works is each of these questions is filled out with additional questions, recommendations and commentary.

This book has the continuous clang! of experience acquired at great expense by a lot of other organizations and hopefully not your organization.

My wife who is not a programmer, but who is a veteran of other communication projects, recognized that this book is what she will need when her employer updates their existing static web site.

The value of this book for a web site builder appears in three places. First it sets the customer up to clarify to you a lot of things about the site before you cut code. Second, there is good advice about domain name acquisition and branding and use of copyright images so the customer can decide to head off domain squatters and traffic thieves before the site attracts attention. Third, there is a lot of clearly expressed technical, html, color and design advice to push you away from obsolete and inappropriate programming practices.

The difference between "Deliver First Class Web Sites" and an early web design book like Greenspun's Philip and Alex's Guide to Web Publishing is ten years of development of the vocabulary and mental landscape that describes a web site.

The Principles of Beautiful Web Design - Design Beautiful Web Sites Using This Step-by-Step Guide, by Jason Beaird, published by SitePoint Pty.Ltd., c. 2007. $39.95 US.

I am reviewing this book because I have done prototype websites with creepy color and ugly graphics. Long ago, I did take art classes and I blended water colors and did drawing projects with the Golden section.

Well, I figure with a little conceptual help I can make some headway in getting my CSS stylesheets and page layouts working better. This looks like the book.

I have found the default page layouts coming from my recent Rails prototypes to have a big problem with scaling gracefully when I enlarge the type sizes using the browser's Text Size menu. Hey, that completely sucks when you are over 40 years old and looking at an old 15" monitor. Anyhow, my next web project has to play well on web enabled cell phones too.

Therefore, this book looks like the third book I am going to use for my next development project.

If you live near San Mateo, California, you may be interested in attending our upcoming PenLUG meeting on Thursday August 23 . The books reviewed here will be at the meeting and available for review.

Meeting title is: Securing Web Applications in the LAMP Environment - Qualys

Meeting URL: http://www.penlug.org/twiki/bin/view/Home/WebHome

The Ubuntu distribution is failing due to not enough high level maintainers.

The failures are amplified because it appears to me that Google has dropped human intervention in the indexing process for Ubuntu or Linux related queries.

And so the observation points right back to me, your typical anonymous user. What should I do, what can I do?

I entered this weekend aiming to restart a Rails program project. I decided to make disk space by copying a couple gigs of jpeg photos to DVD. So I was going to install a $25 Emprex brand DVD writer, a sale item from Fry's.

The DVD writer inside the box was actually a Sony brand drive. It reads and writes CDs fine. Eventually I determined that a dvd writing application called growisofs was not working right.

So at that point I had a plain application software problem.

One step I took was to search Google for solution.

Google isn't returning helpful results as it seemed to do in the past. In the past I have used Google to search for solutions to software problems. I have had the feeling that I was getting excellent results because somebody had edited the search results to force the "answer" to show up at the top of the search results.

Another step I took was to search the Ubuntu forums for a solution.

Ubuntu forums has lots of earnest people but important information about application failures isn't reaching the responders.

By exerience, I find I have to search the forums carefully for well phrased questions and quality answers. It turns out that in the Ubuntu forums, "DVD" is the last item in the most searched for word list. Going straight to the "DVD" item, there was a very reasonable 250 forum posts, so I reviewed the ten pages of posts and read the most plausible 20 support exchanges for my particular problem.

Somewhat against my instincts, the most plausible Ubuntu forum posts said that problems like mine were best solved by doing a distribution upgrade. I have been sticking with "Dapper 6.06 LTS" and the posters recommended moving to "Feisty 7.04".

Lacking a quality answer about the specific application I know is failing for apparent lack of memory, I reluctantly accepted the low quality answer of "do a distribution upgrade". The upgrade is stuck at a character terminal and I'll spare you the gory details. I would laugh except it hurts too much.

So it looks to me like those few elite programmers who intimately understand the Ubuntu distribution are not involved enough with the failures that accumulate as the kernel evolves and applications and hardware details fall behind.

I can't say with my bare face hanging out that Ubuntu is a productive environment for me at this time.

Mini review: Practical Rails Social Networking Sites by Alan Bradburne, published by Apress, Berkeley, California 2007. US $44.99

Reviewed by Lee McK

Practical Rails Social Networking Sites has a cover subtitle "Learn how to implement a modern social networking web site using Rails, from design to deployment."

This book walks the reader through the steps of setting up Ruby and Rails software on a Linux or Windows (or BSD or Apple) computer and then designing and setting up several of the well known modules that appear in social networking websites like myspace.com and pbwiki.com.

The skill level expected by this book is you have worked through several chapters of an introductory Rails programming book. The back cover names two Apress Books as an introduction: Beginning Rails: From Novice to Professional. Also from Apress, Beginning Ruby: From Novice to Professional.

A book I used that prepared me for Practical Rails is Agile Web Development with Rails by Dave Thomas and David Heinemeier Hansson. I probably invested 4 weeks or 50 hours doing the tutorials and then building two prototype web sites. (How much time I admit to spending depends on whether my wife is listening.)

I think this book is a really good tool if you have explored Rails and your aim is to develop a database backed website with users, blogs, email, user created pages, user uploaded photos, and use of published API services such as Google Earth.

My experience so far with Rails kind of came to a stop when I built a prototype website using a Gem called Goldberg (I think that is the name). This gem made it easy to create the frames and pages I wanted. So I jumped immediately to the core of my problem which required components for uploading global positioning system data and displaying the data on topographic maps. I got bogged down with geographical information systems problems and used up my development time. I shelved the project. Practical Rails: Social Networking Sites may be the casebook of examples that will help me restart this project.

Now this is where Beginning Rails: From Novice to Professional. comes in as a guide. The author Alan Bradburne has divided the social networking website problem into about 10 components. For each component he walks through the classic Rails module development steps: you specify the requirements, sometimes you download a gem, then you create the model, the controller, the view and the tests.

This book is partly recipes. The website described in the book appears at http://railscoders.net

This book is in larger part an experienced developer showing the beautiful and powerful Rails "model view controller test" program design discipline.

For instance, one component is building a photo gallery. The author shows the classic Rails development steps as you modify the data model, you add processing in a controller and add display layouts in views and finally you write tests so you can promptly find out if later changes break earlier code.

07/25/07

I am puzzled at how my experience upgrading my Ubuntu distribution is not being reflected in reviews or news.

My experience as a long time Linux user is: Ubuntu 6.06 LTS has enough configuration errors that it shouldn't be offered as an end user distribution yet.

I upgraded from the 5.05 "Breezy Badger" Ubuntu to the 6.06 "LTS" about 6 weeks after LTS was released. After a news report mentioning a bug with the X windows driver I waited 3 weeks. Nope X was still broken.

The xserver package was still broken. It immediately broke my X display. From the ubuntu forums I got instructions to push the xserver package version back. Four hours later I got my X display back.

What has happened since then is one little application after another is coming up broken. Mostly minor changes are breaking applicatons. I have not found a single major improvement in 6.06 that compensates for the 16 hours wasted. I haven't messed with ticky tacky stuff like this since the RedHat C-compiler version problem of 5 years ago.

The agony is I am a bloody desktop user doing Craigslist job search and Rails. I actually need Google Earth to "just work".

I really am the wrong guy to back up and futz with problems like "the openGL graphics functions for the mga driver seem to be broken." For a long time now, I have depended on the Debian package system to set things up right.

The upgrade water torture ticky tack list for LTS:

- Local printing broken - required delete and reinstall printers.
- Adobe acrobat no longer being called as a firefox plugin (still broken despite messing with Preferences)
- Google earth now showing OpenGL error message (still broken)
- Matrox mga video driver, can't tell if the opengl functions needed by Google earth are working.
- mplayer stopped, requires an obscure edit to /etc/mplayer.config
- sound for YouTube videos now requires opening a new alsa package and futzing with mute settings. Desktop sounds are gone. (configuration for alsa-mixer changed).
- Firefox closes all windows (or crashes) when certain web pages are closed by the server.