Posted by Nick Sieger Sat, 21 Oct 2006 03:22:00 GMT
Ok, ok, so I’ve been caught redHanded. Yeah I know, Tufte hates bullets. You might feel a strange sensation if you stare at my transcriptions of today too long. If anyone is offended or would like to claim I’m a card-carrying NRA whipping boy, let me know and I’ll try to tone it down a bit. Thanks for the mirror of truth, _why!
Posted by Nick Sieger Sat, 21 Oct 2006 02:59:38 GMT
An annual tradition at RubyConf is the “Roundtable”, where any member of the audience can come up and ask Matz any question. Transcript follows with partial paraphrasing.
Q. Will Ruby ever get the features of evil built in? Will we get become? No.
Q. If Symbol is a “frozen” String, why do we need Symbol? I don’t have a complete answer to that question – following tradition.
Q. What’s the most unusual architecture Ruby has ever run on? Ruby…I can’t think of one…JRuby? Some guys ran compiled Ruby on the NECS supercomputer. [From the audience] Symbian.
Q. As Ruby gains acceptance, it becomes more resistance to change. How do you keep agile while gaining momentum? We have forked off 1.8, so if you want stable, use Ruby 1.8 forever.
Q. The array patch – is there any chance of a backport for those of us who want 1.8 forever? Could be, but the current patch has a bug in 1.9.
Q. Why isn’t YARV the only VM for 1.9? [Comment about register-based VM] I’m not sure whether stack-based or register-based VMs are better…[THAT GUY!]…[Given the hook]
Q. It seems like momentum is based on strong metaprogramming facilities. With 1.9 there are more user-friendly features. Will code eventually be data, or are we just patching on techniques on a tool that is being used more heavily? An extreme way to do metaprogramming is to use Lisp, but for practicality we have to start somewhere. I’m not sure that s-expressions are the way to go. We’re not going to have macros. It will remain similar to what we have in 1.9. The ability to serialize code would be really powerful – will that happen? Disclosing internal state is tricky because it changes. It could be a platform-specific feature.
Q. Recently for 1.9, there was a patch for Array and String to be in the object header. Was this profiled? It seems like it would be an incredible performance problem. It was done for slowness of malloc and gc. In some cases it could run slower, but it benchmarked at about 5% faster. It should be tested thoroughly.
Q. Ruby 2.0 will not have green threads or continuations, correct? Are there specific reasons for that decision? It is difficult to implement green threads in YARV.
Q. People want to see development move quickly, but may not be able to contribute code. Can you suggest ways for those programmers to contribute? Submit patches to the list. What is the obstacle to joining us? [Question restated] Submit testing, submit documentation. Contribute to RubyCentral? (looks at Chad Fowler)
Q. The CSV library is useful, but slow. Can we see FasterCSV in stdlib? It’s ok to replace the csv library with FasterCSV, as long as the compatibility issues are resolved.
Q. I want to thank you because I love writing in Ruby every day. I tried to write Python first. Recently an unrelated upgrade broke my Python apps. Will we be able to manage multiple versions of Ruby? You’ll have to have two versions, 1.8, and future ones.
Q. What would you say is the most important feature of your personality that has given you success? Endurance? Can you elaborate? It’s easy to design a language, many of them disappear in a year or two. I’ve been working on Ruby for 13 years.
Q. I wanted to see in IRB how the class was implemented at runtime. (He means an uneval feature) [Audience member (Eric Hodel?) mentioned ParseTree]
Q. If there wasn’t Ruby, what would you be programming in? Some language of my own not named Ruby.
Q. Do you still look at RCRchive? Yes. I’ll talk about it in the keynote.
Q. I would like a feature abstraction that would allow Ruby to fork itself rather than actually forking. Can you see this added to core? We have to define the behavior of that feature. We have to have a good name for the method. Then there will be no problem to add it to core.
Q. Why do you need to bind methods to an object of the same original module? If you bind a method taken from String class to Array class, it will crash, and we need to prevent it.
Q. What do you think of Ruby.NET? I’m pretty open to new implementations. Should there be a standard (language specs) that other projects should follow? Charles started work for a written 1.8 spec, and I’ll help him with that if I have time.
Posted by Nick Sieger Sat, 21 Oct 2006 02:54:00 GMT
gem install radiantmkdir demo && cd demo && radiant .cp config/database.sqlite.yml config/database.yml<r:title/>, <r:content part="sidebar" inherit="true"/>, <r:snippet name="footer"/>, <r:if_content part="extended">...</r:if_content>, <r:children:each limit="5" order="desc">...</r:children:each><r:title/> picks the correct title even if it is embedded within a snippet within multiple pages class MyFilter < TextFilter::Base
register 'profanity filter'
def filter(txt)
txt.gsub(/(damn|ass|shit)/i, '####')
end
end
Radiant is powering the new ruby-lang.org site. Overall, a slick, well-thought out, polished, extendable CMS done very much in the philosophy of Ruby and Rails. Check it out!
Posted by Nick Sieger Sat, 21 Oct 2006 02:51:38 GMT
net/httpPosted by Nick Sieger Fri, 20 Oct 2006 20:18:00 GMT
Geoff Grosenbach is waxing on pagefuls of numbers condensed into a small, tidy graph that increases the amount of information you can communicate on a page.
The cartoon fox makes his second appearance of the day in one of Geoff’s sparklines
It is in our hands, the hands of the programmer, to show designers what the range of visual representation capabilities are.
caches_page) an image with textresponds_to {|type| type.jpg { ... } }Posted by Nick Sieger Fri, 20 Oct 2006 19:07:00 GMT
Update: Evan has posted code and has a page set up for the project.
Evan Phoenix (nee Webb), of Seattle.rb, is presenting on Sydney and Rubinius, an experiment in improving the ruby interpreter. Sydney has died, and Rubinius has risen from its ashes, appropriately.
Today’s Ruby interpreter is like a big dump truck – sometimes a little slow, but it works for us. YARV is like the red, shiny fire truck. Both big and complex. Rubinius, by comparison, is like a dune buggy. Fast, light, but you’re going to get sand in your eyes if you drive it a lot.
The project, admittedly, is naive.
Q. Since you were starting over, could you use a platform-independent library to ease the process, such as APR? Yes – currently using String and PointerArray from glib.
Q. How is performance? Too early to tell – I hope to know by the end of the conference. Prototype S became runnable and usable on the plane here.
Q. Can you clarify the goal? To create a Ruby interpreter in Ruby that can translate itself out into a C interpreter.
Q. Have you figured out how to link in external libraries in a platform independent way? No. My hope is that the decision will be made to write a common framework for translating to system calls, e.g., SWT.
Q. Have you looked at PyPy? (similar project for Python) Yes, and it’s f-in complicated. It worries me actually.
Q. Could you have it generate backend code in another language/platform (Java bytecode, CLR)? Yes, I certainly hope so, otherwise I’m wasting my time.
Q. How will you add native thread support in a cross-platform way? I hope I won’t have to, by leveraging external tools.
Q. If you’re building a Ruby-to-C translator, why write a Ruby interpreter at all? If I didn’t, what would I translate? You still need some core engine to translate. Would it be a subset of Ruby? Yes.
Q. Looks very similar to Squeak, have you looked at Squeak code and talked to Squeak people? Looked at code a lot, I’ve really stolen all of their ideas. I haven’t talked to the folks yet because I’m afraid they might laugh at me.
Q. Worried about fragmentation? Yes, but I really want to make it as compatible as possible with the current interpreter.
Q. Rubinius bytecode compatibile with YARV? No, but I hope to be able to write a bridge to YARV in Rubinius.
Q. Have you looked at Valgrind for the C code? Yes, I have. Good possibility for future direction.
Q. Can you demo some code? They’re incredibly boring. “Look I got a MethodTable object, I asked for one.”
Posted by Nick Sieger Fri, 20 Oct 2006 19:06:00 GMT
Takahashi-san is here to present on the history of Ruby, an apparently thankless task, because none of the other original Rubyists are historians. Takahashi is the co-author of two Japanese books on Ruby, Enjoy Ruby and Ruby Recipe Book. He also has the “Takahashi method” of presentation named after him. His talk presented an informative timeline of Ruby, the details of which were a bit tricky to capture. If I transcribed anything erroneously, please let me know.
Posted by Nick Sieger Fri, 20 Oct 2006 19:04:00 GMT
The RubyConf room is filling up this morning. I’ll be doing my best to live-blog the conference here so stay tuned!
Posted by Nick Sieger Fri, 13 Oct 2006 03:20:00 GMT
My name is Nick Sieger, and I’m an Emacs-o-holic. I’ve been using Emacs for over a decade, and my .emacs file alone (excluding
custom elisp packages) has accumulated 125KB of cruft. Basic navigation keystrokes are beyond ingrained into my fingers, they’re
automatic. I was pleased when I found Jacob Rus’ explanation of the Cocoa text system and accompanying instructions for baking
these commands into the basic OS X text widget; now getting around Cocoa text widgets is a breeze.
Emacs, in a way, has a reputation for being something like the Spruce Goose of computer programming tools. Some view it as the pinnacle of achievement in text editing tools; a piece of art; the most extensible piece of software ever written. On the other hand, the learning curve is considered to be much steeper, and Emacs hasn’t changed visually very much over the years, sticking to a minimalist interface that doesn’t dazzle like tabby, tree-widgety, multi-paneled IDEs.
So it seems that I’m going to be continually comparing TextMate to Emacs; perhaps a subject worth expounding upon. As I ramp up on TextMate, maybe these findings and tips will become useful for anyone, Emacs user or not.
The “project” system is simple but elegant in its conception. Open a directory, and it becomes your workspace. Navigate to any file quickly with ⌘-T. I always wanted something like this in Emacs and considered writing one at one point.
Visually appealing. Renders fonts well, beautiful built-in color schemes, strong syntax-highlighting. The multiple, nested context highlighting is well done. At the same time, the UI is minimalist enough that you’re not distracted by gratuitous toolbars full of icons that you constantly have to hover for the tooltip to remember what they do.
Feels right. The app blends in well with the OS X user interface, and shares its philosophical underpinnings of simplicity and least surprise.
Broad language support. Many popular languages supported out of the box (notable exception: C# – I wonder why), and if you get subversion bundles, many less common ones, e.g., Io, Lua, or OCaml (note to self: none of which I have had the chance to explore yet).
Community support and innovation. Seeing the buzz in the community around TextMate is strong evidence that it’s spurring some innovation, a fresh approach, in a category of software that hasn’t seen much for quite some time.
Text selection while staying on the home row. Setting the mark (C-SPC) and moving the cursor to select a region of text. I don’t like moving my right hand over to the arrow keys to make a selection! This is my number one new feature wish for TextMate.
Buffer selection with C-x b or C-x C-b to quickly jump to an already-open file. ⌘-T replaces this to some extent, but it’s not the same and requires some retraining. Update: ⌘-T RET is equivalent to C-x b. Good enough. One thing I have discovered that I like quite a bit is the ability to scroll through the buffer/file-selection box with C-n and C-p.
back-to-indentation. This is an emacs command that moves the cursor in front of the first non-whitespace character on the line and is
bound to M-m by default. This keystroke is apparently one that I use frequently, because my fingers keep typing ⌘-M which has
the effect of minimizing the current window in TextMate. Ack.
open-line in Emacs is bound to C-o, which inserts a new line where the cursor is but leaves the cursor before it. The exact same
command is in TextMate, and is bound to the same key. Sweet! But apparently my fingers are trained on a variation: split-line, bound
to C-M-o. This is similar to open-line except that it moves any text to the right of the cursor vertically down. When you’re sitting
at the first non-whitespace character on a line and you split it, it has the effect of creating a new line directly above the current
one, leaving you at the same indentation level. Yes, C-a C-o TAB pretty much does the same thing, but it’s three keystrokes instead of
one.
hippie-expand is a powerful, extensible multi-completion mechanism that is hard to live without. TextMate’s ⎋ completion seems
roughly equivalent to Emacs’ dabbrev-expand, but hippie can expand similar words and lines from all open buffers, making the
completion that much more powerful and context-aware.
recenter (C-l). Puts the current line in the middle of the window, unconditionally. TextMate doesn’t appear to let you scroll the
contents of the window such that the last line in the buffer is any higher than the bottom of the window. I prefer to have the line of
code I’m editing closer to the center of the window.
string-rectangle. This versatile command allows you to replace each line in a rectangular selection with a prompted string value.
Great for adding a column of tab characters, commas or other formatting to a group of similar lines. Update: I found the TextMate
equivalent for this one. It turns out TextMate has a great rectangular editing facility, simply by starting a selection, tapping
⌥, and continuing. Typing while the rectangle is active inserts new text in the same manner as string-rectangle. Excellent.
What else have you found about TextMate that reminds you of Emacs or helps you get over it?
To be continued!
Posted by Nick Sieger Wed, 13 Sep 2006 20:35:00 GMT
Update: (2 months later) If you’re reading this, you’re probably interested in my Rails plugin for this instead.
Hot off the presses, after a few hours of hacking and tweaking, may I present Auto+RSpec, otherwise known as The Mashup of RSpec on Rails and autotest. This is not an official release of any sort, but “may work for you.” It’s not a clean hack, as it exposes some areas for autotest to grow if the maintainers decide to open it up to alternatives to Test::Unit. After spending a little time looking at the autotest code, I think it would be nice to allow hooks for autotest plugins to define project conventions (i.e., @exceptions and the #tests_for_file method) as well as a result parsing API.
For now, if you’re an RSpec on Rails user, you can try this out as follows:
sudo gem install ZenTest.vendor/plugins/rspec/lib directory (you did say you’re using RSpec on Rails didn’t you?)lib/tasks directoryautotest with rake by typing rake spec:autotestNext steps for this will be to work out whether this code should live in RSpec on Rails or autotest, or some combination of those.
Now, spec’ers, be off in search of that Red/Green/Refactor rhythm of which sage agilists speak!
Bonus tip: add the following code to your .autotest file to run spec with rcov:
Autotest.add_hook :initialize do |at|
# run spec with rcov
if at.respond_to? :spec_command
at.spec_command = %{rcov --exclude "lib/spec/.*" -Ilib --rails "/usr/lib/ruby/gems/1.8/gems/rspec-0.6.0/bin/spec" -- --diff}
end
end