DISQUS

Python:

print "Hello, world. " * 10

reply

Morris

Lecture 2:
Why the Array#inject example, for inserting only doubled values of even numbers, failed:

The problem is that "expr1 if expr2" returns the value of expr1 iff expr2 is true, but if expr2 is false, the value of the whole expression is nil, hence the next time through the block, the memo has the value nil. Instead of "; m", a way around this as a more-pure one-liner would be to use the block

{ |m, e| (e % 2 == 0) ? (m << e*2) : m }

guaranteeing, via the ternary operator, that the expression has the correct value (the accumulated memo) whether or not e is even.

reply

xertroyt

Very good. That would also give you the freedom to squeeze in a semi-colon for another purpose, though our "semi-colon exception" is exactly for this usage.

reply

Mario

Student Locations Google Map

I'm looking at the shared map, and I can't see the map pin icon . It looks like everyone has 'Edit' access. Is it possible that someone has deleted the map pin icon by mistake, or am I just blind?
--MC

reply

xertroyt

Morris reported that you need to be logged in to Google first. So if you have a GMail account, make sure you're logged in.

reply

Mario

nope - logged in and I still can't see anything. That's OK - there's enough icons near Harvard Square anyway. --MC

reply

Keith

Hmm. If you click 'edit' there should be four icons in the upper-left corner of the map itself, in case anyone else is having problems.

reply

Morris

The book _Advanced_Rails_ , recommended by John for chapter 1 about Metaprogramming, is available online through the Harvard Libraries' subscription to the Safari bookshelf, free to anyone with a Harvard ID and PIN. Just go to http://holliscatalog.harvard.edu/ , search for the book, and click on the "INTERNET LINK". The Flanagan and Matsumoto book _The_Ruby_Programming_Language_ is also available, and, presumably most (all?) other O'Reilly books.

reply

Morris

In last week's lecture (which I've just finished watching), someone asked whether an eigenclass could have class variables. John said no, but a little experimentation in irb shows that the answer is yes. I'm not sure why you'd want them, but they're there (and, as far as I can tell, should only be available through the eigenclass's methods):

foo = "I am a string"
def foo.method1
@@eigenvariable = 'I am a class variable'
end
def foo.method2
puts @@eigenvariable
end

foo.method2
NameError: uninitialized class variable @@eigenvariable in Object

foo.method1
foo.method2
I am a class variable

reply

xertroyt

This is a case of Ruby trying to use the same rules for classes and objects even where they don't make much sense (at least they don't make much sense to me). I guess you could imagine wanting a private namespace for the eigenclass, but . . .

reply

Student1

I just happen to need to compare 2 files. Actually 2 directory of files.
So I thought it might be a good opportunities to try doing this with Ruby.

RDoc says require 'ftools'. Do I just do it with "gem install ftools"?
Normally how do people get more packages and is it safe to just keep adding packages in ?
I don't want to mess up my environment for other assignments to come.
(I am assuming that I have to add ftools in by doing gem install)

Is there a way to find out what other things we can install.
I did a gem -list, but I assume that lists something I have installed.

On a separate question, does something like methodA(self) work ?
I mean can one pass self as a parameter.

Thanks

reply

Keith

ftools isn't a gem, it is part of the standard library:
http://ruby-doc.org/stdlib/

require 'ftools' in your script should work fine; you can always check in irb, if you type 'require 'ftools'' and true is returned, you have the library available.

Also: self is an object, and as such can be passed as a parameter. You would more often see 'yield self' rather than seeing self passed as a method parameter, I think.

reply

Student1

Thanks.
Yes, I have ftools.

reply

philadelphia

Hi,

I'm working in Eclipse with an Aptana plugin... Running on Windows and Linux, on two different machines. Sqlite3 works just fine, but I'm trying to use my Derby DB, which is bundled with Elipse for development. I've update the .yml file to point to the jdbc driver required to access the Derby DB, but I'm having a tough time working with Derby. The immediate problem is when I run "ruby script/generate scaffold test_db title:string" I get the following error:

/usr/lib/ruby/gems/1.8/gems/activerecord-2.1.1/lib/active_record/connection_adapters/abstract/connection_specification.rb:232:in `establish_connection': Please install the jdbc adapter: `gem install activerecord-jdbc-adapter` (no such file to load -- java) (RuntimeError)

I've installed the activerecord-jdbc-adapter using gem... Not sure what I need to do. I figure it's something really obvious I'm missing. Any help is appreciated.

Thanks,
Mike

reply

xertroyt

I think that AR-JDBC bridge is for *JRuby*.

But you are typing:

ruby script/generate scaffold . . .

(i.e., ruby.)

Do you want ruby or jruby?

I haven't used JRuby with rails, but try

jruby script/generate scaffold . . .

Also: This is way outside the scope of the course. If you want a recommendation, I would say: Write your app with the Matz interpreter (e..g, ruby) and then when it's working, try GoldSpike or one of the other strategies to get your app to run inside a JavaEE container.

reply

philadelphia

OK... Thanks. No worries. Will continue to fool around with it... Just creating a bunch of sample projects so that I can make sure I understand all of the steps required and how all of the MVC components ( and others) work together. The external DB was part of my experimentation with the yml file.

Not critical to understand at this moment. Will figure it out eventually.

Thanks again,
Mike

reply

xertroyt

My gut says: Stop using Eclipse/Aptana right now. Learn Rails with an ordinary text editor.

At present you are trying to change three things at once: The implementation of Ruby itself (from Ruby to JRuby); the runtime (managed by the IDE); and the database.

reply

katemuse

Wouldn't it be nice if rake spot/test printed the errors in reverse order? Searching up the terminal window for the first error in a long string of tests is a time consuming pain. I wonder if anyone has ever tried printing tests to a file, reversing the paragraph order and then reading the file from the terminal.

reply

xertroyt

There are test runners that present a GUI. But at this point, it is best to struggle with the text output. Consider it an incentive to get your code to pass more tests (and reduce the output).

reply

katemuse

Along the same lines: wouldn't it be nice if these comments appeared latest-first?

reply

xertroyt

Good idea -- I just switched it.

reply

katemuse

cool!

reply

Mario

I'm sorry if this was mentioned somewhere and I missed it, but I attempted to mail my ZIP package to Keith at Gmail.com , and they don't take ZIP or TAR.GZ files. I'm including the error message here as evidence. -- MC

209.85.217.181 failed after I sent the message.
Remote host said: 552-5.7.0 Our system detected an illegal attachment on your message.
552-5.7.0 Please visit http://mail.google.com/support/bin/answer.py?an...
552 5.7.0 to review our attachment guidelines. 6si9929189gxk.79

reply

xertroyt

They most certainly do take .ZIP files.

The error is that you left the PKZIPC.exe file in the bin/ directory.

I said in e-mail and in posts: remove this file first.

John

reply

akv

have the assignment grades been distributed yet? The email said early thursday morning or wednesday.

reply

xertroyt

I asked the TAs last night to e-mail the grades.

reply

nocsav

I keep thinking about the idea of implementing constraints inside the DBMS vs. Rails. Wouldn't people who have more than one kind of program modifying their database benefit from doing constraints inside the DB?

reply

xertroyt

Yes, absolutely.

If you know the syntax for how to add constraints in your database, you can run raw SQL through AR to define those constraints -- I think I give the page number in my slides.

Which reminds me, I need to post the slides! Will do so tonight (slides are on my laptop, which I don't have with me at work today).

reply

philadelphia

Hi,

If I want to use Ruby/Rails to access an existing DB, and its naming convention for the tables is inconsistent with that of Ruby/Rails (e.g. single/plural), how do I set up the ActiveRecord:Base extension? For example, if a SQLITE3 table is called ABC_123, what does the model class look like?

Thanks in advance,
Mike

reply

xertroyt

Mike,

Take a look at p. 296 in the AWDR PDF. You can say "set_table_name" in the Model to bind a model to any table you want.

John

reply

philadelphia

Great! Thanks very much.

Mike

reply

Morris

During last week's lecture (lecture #7?), I asked about what gets saved when, if an object is added to a "has_many" collection, e.g.

class User < ActiveRecord::Base
has_many :stocks
end

class Stock < ActiveRecord::Base
belongs_to :user
end

u = User.new()
s = Stock.new()
u.stocks << s

The question was when does this relationship get written to the database?
Being somewhat behind in my reading, I just found this in the textbook, in material that should have been read before the lecture. Page 372 of AWDR, the last part of section 18.7 "When Things Get Saved", and the answer is "it depends". If the parent object (in this case, User u) exists in the database, then adding the child object (in this case, stock S) to the parent's collection automatically saves the child (i.e. updates the child's database row). If the parent is NOT yet in the database, then the child object is not saved until the parent is saved.

reply

xertroyt

Thanks. Here's a bit of sample code (for migrations2) which demos the case -- followed by the output SQL.

require File.expand_path('setup')
newguy = User.new(:email => "neville@example.com", :first_name => "Neville")
internap = Stock.new(:company => 'Internap', :symbol => 'INAP')
newguy.save!
newguy.stocks << internap

jgnmbair:migrations2 jgn$ ruby demo4.rb
User Create (0.001168) INSERT INTO "users" ("updated_at", "phone_number", "first_name", "password", "last_name", "email", "created_at") VALUES('2008-11-03 07:03:43', NULL, 'Neville', NULL, NULL, 'neville@example.com', '2008-11-03 07:03:43')
SQL (0.000460) SELECT "company" FROM "stocks" WHERE ("stocks".company = 'Internap')
SQL (0.000183) SELECT "symbol" FROM "stocks" WHERE ("stocks".symbol = 'INAP')
Stock Create (0.000524) INSERT INTO "stocks" ("company", "updated_at", "symbol", "user_id", "created_at") VALUES('Internap', '2008-11-03 07:03:43', 'INAP', 3, '2008-11-03 07:03:43')

reply

Morris

The partial crud example code, from lecture 9, contains calls to the form_for helper method, where part of the parameter list is

:url => { :action => :update, :id => @section }

and when this gets processed by erb, the action URL contains @section.id at the end of it, just as if we had written

:url => { :action => :update, :id => @section.id }

except that this works even if @section is nil, omitting the id in that case.

Playing with this in my own code, I see that this works the way one would like it to, eliminating the exception which would get raised when @section is nil.

But where is this documented? I can't find it in the on-line Rails documentation, and I suspect that it's somewhere in AWDR, but I can't locate it there either.

reply

xertroyt

I can't reproduce that -- maybe I'm misunderstanding the problem. You mean even though @section is nil, when you take @section.id in the view you don't get an exception?

E.g., if in SectionsController I do this:

def edit
@section = nil
end

and then in views/sections/edit.html.erb do:

<% form_for :section, :url => { :action => :update, :id => @section.id } do |form| %>

I do get the nil exception ( "Called id for nil, which would mistakenly be 4 -- if you really wanted the id of nil, use object_id" )

reply

Morris

There's nothing to reproduce, and it's not a problem. (Mis?)reading the documentation, I would expect that the :id key would take an id number as its value, so that the code (if @section is not nil) should be
:url => {:id => @section.id}
and in fact, this works exactly as I expected. And if @section is nil, then we get the exception, also as expected. No surprises here.

BUT, we can also just use @section as the value for the :id key, and, as with so much in Rails, magically the right thing happens: apparently, Rails detects that the value is not an integer but instead an ActiveRecord object, and calls its id method (and knows enough NOT to call the id method when its argument is an id number). And if @section is nil (or maybe it's if @section.id is nil), the id number is not added to the end of the URL. This allows us to use the same partial template, _section.html.erb, for edit_with_partial.html.erb, where @section = Section.find(id),
and for new_with_partial.html.erb, where @section = Section.new (a Section object with a nil id, not, as I thought before thinking it through, nil).

I don't remember for sure what error I was getting, but I'm pretty sure that the error message said something about a nil. Unfortunately, I can no longer reproduce this error. I was probably mistakenly passing in a nil somewhere where I shouldn't have been, or had a name mismatch, or something else which a more strongly-typed and declarative language would have caught before getting that far.

But my question remains: for the :id key in the :url hash, we can provide an id number, OR we can provide an object which has an id attribute, and in either case Rails does what we would like. The question is where is this documented? Other than copying your code from the lecture examples, how would I have known that I could use either an ActiveRecord object or an integer here? Where in AWDR or the on-line Rails doc should I be looking, to understand this a little better instead of just treating it like magic?

reply

xertroyt

Sure, the use of an AR object without explicitly using the .id attribute is documented in AWDR - see in a bit.

I'm more concerned about the nil error you were getting -- that's what concerned me.

In any case, I think this is what you're looking for:

AWDR, Sec. 20.2 (p. 417 in the latest PDF):

Create a URL that references this application
url_for(option => value, ...)
Creates a URL that references a controller in this application. The options hash supplies
parameter names and their values that are used to ﬁll in the URL (based on a mapping).
The parameter values must match any constraints imposed by the mapping that is
used. Certain parameter names, listed in the Options: section that follows, are reserved
and are used to ﬁll in the nonpath part of the URL. If you use an Active Record model
object as a value in url_for (or any related method), that object’s database id will be used.
The two redirect calls in the following code fragment have an identical ef fect.
user = User.find_by_name("dave thomas")
redirect_to(:action => 'delete', :id => user.id)
# can be written as
redirect_to(:action => 'delete', :id => user)

Unfortunately, I don't think it's documented in the RDoc for url_for, routing, or the other places I might expect it.

reply

Morris

Thanks. What a (relatively) obscure place for it.
I'm fairly sure that my nil error crept in while I was refactoring my code to use a partial template, and the partial template must have been using some variable that didn't (yet?) have a defined value within the partial.
Somehow, in the course of fixing the wrong problem, I accidentally fixed the right problem.

reply

xertroyt

Well, I'm going to be going into the routing system in tonight's lecture, and will be talking exactly about some of that magic.

reply

philadelphia

Hi,

In the example online, within the show, used to create a new observation, you have the metric kind (which I believe is the observation_kind.name) displayed within the title area.

But in the static HTML distributed, you have this information appearing outside of the title.

I'm guessing you meant to have it in the title. But I want to make sure this is not a distinction you made intentionally.

Thanks,
Mike

reply

xertroyt

When you add an observation here: http://metricsmine.plugh.org/observation_sets/s...

you should see the name of the observation set in two places:

(1) In the <title>
(2) above the form (e.g., Metric: my weight)

Is that what you mean?

So, for:

(1) The <title> bit is set *in the layout*: You have to set the right variable in your template to make this work. This pattern is also used in CCC, LinkWizz, etc.

(2) should be set with your code.

Take a look at the welcome.html.erb to see how the title is set -- and then look at the layout.

John

reply

Morris

What philadelphia Mike is referring to is that, unlike in most of the static HTML examples, and unlike the on-line reference implementation, the sample static HTML for observations_sets/show has the title just as a plain text above the form, rather than in an <h1> element. I agree with Mike, in that I think this was probably unintentional in the static example.

reply

xertroyt

I must not be seeing what you're seeing. Do you mean THIS in the static HTML for show.html.erb ?

<%# Next line should show metric kind dynamically; if
# it's null, show nothing.
%>
Metric kind: personal weight

reply

philadelphia

Yes... the Metric kind: personal weight.

It appears, in the static code, outside the title area (bracketed by h1's above it).

And in the online version it does not exist.

Mike

reply

xertroyt

That's different from the title.

Log in as: john
Then go here:
http://metricsmine.plugh.org/observation_sets/s...

The title is: Metric: calories (raw data)

This is in two places: In the document's <title> tag, *AND* inside of h1's like so: <h1>Metric: calories (raw data)</h1>

THEN just a bit before that, we give the METRIC KIND, if known:

Metric kind: CollegeFund - test1

What has happened is that after a server restart, someone deleted the observation_kind for that observation_set, so it isn't showing for "my weight," etc.

reply

Morris

Yes, exactly what Mike said. And in the reference implementation, it's expressed as "Metric: whatever" rather than "Metric kind: whatever", and the contents of the <h1> element matches the page title.
This (matching <head>/<title> and <h1>) seems more correct than just plain text outside of any HTML tags, and is stylistically consistent with the rest of the site.

reply

xertroyt

There are simply no <h1> tags around the indication of the metric kind. It is not intended to be consistent. The only part that is consistent is the main title tag, and the <h1> repetition of what is inside of the title tag. The indication of the "metric kind," i.e., the observation kind for the particular observation set, is EXTRA. The name of the METRIC is prefixed with "Metric:" and the name of the METRIC KIND (i.e., observation kind), if known, is prefixed with "Metric kind:". It is different from the name of the observation set. It's simply a little bit more information about the data in question.

So: In view/source from the reference implementation (much whitespace removed):

<title>MetricsMine - Metric: calories (raw data)</title>
<div id="main">
<h1>Metric: calories (raw data)</h1>

Metric kind: CollegeFund - test1
<form action="/observation_sets/create_observation" method="post">

In the view template stub (the static HTML) for observation_sets/show.html.erb:

<center>[PARTIALLY DYNAMIC - PLEASE IMPLEMENT!]</center>

<%# Next line should show metric kind dynamically; if
# it's null, show nothing.
%>
Metric kind: personal weight
<form action="/observation_sets/create_observation" method="post">

Remember that the <h1>Metric: calories (raw data)</h1> is not in the view templates -- it's in the layout.

reply

Morris

Got it. I think your previous explanation was clear. Our confusion
was arising from the fact that none of the pages being presented by
the on-line version happened to be showing this particular item. I
will have to cogitate a bit to figure out under what circumstances
this item would be non-existent.

reply

xertroyt

There is a page where you see it: the "calories" set that someone created.

What I can going to do is reset the database so you can log in and see it with the "canned" obervation sets.

The app will be down for a bit, and this may cause havoc with your cookies; you may find that you have to log out, delete your cookies for metricsmine.plugh.org and/or close / reopen your browser.

When an obsrvation set is created, it is tied to a kind. The idea is that if Morris creates a set called "my weight" and I create a set called "my weight," they might both have the kind "personal weight." So that way our data can be compared: They are both of the same kind.

However, we allow kinds to be deleted. When this happens, the association from the set to the kind will be null, and there will be nothing to display.

reply

xertroyt

OK, guys.

First, definitely clear your cookies, close your browser, and come back to metricsmine -

Then log in as "john" - and look at http://metricsmine.plugh.org/observation_sets/s...

Now do you see it? We show the "kind" for the set, if it is available.

reply

xertroyt

If you will PLEASE LOOK AT THE STATIC HTML, you will see that it says:

<%# Next line should show metric kind dynamically; if
# it's null, show nothing.
%>

IF IT'S NULL, SHOW NOTHING.

So: If you will log in as admin on the reference implementation, notice that someone has DELETED the original observation kind from which the observation set was created. THEREFORE, the kind is empty. THEREFORE, there is nothing to show. THEREFORE, in the live version, nothing is shown.

THEREFORE, to help you out, I put in an explicit comment in the static HTML that is telling you to dynamically show the OBSERVATION KIND if it is available.

reply

philadelphia

Hey John,

This is the kind of thing that would be a lot easier to ask in person. Something like, "hey, is this supposed to be there?" With a response like "yup".

When taking a course at a distance, you can't just point and ask a quick question. You need to spell it all out - in an open forum. Which appears to make questions much "bigger" than they are.

Thank you for the answer,
Mike

reply

xertroyt

You are right about it being easier in person -- I have taught other subjects on-line (English literature at Ohio State; media studies at the New School in NYC) -- and the distance / computer-mediation produces all kinds of communicative problems. It may be that technical topics like engineering and CS are especially difficult online because of the impedance mismatch between computer and human languages.

In any case, the discussion started with: "In the example online, within the show, used to create a new observation, you have the metric kind (which I believe is the observation_kind.name) displayed within the title area." So that took us down a tricky path, because the kind is not shown in the title -- only the name of the observation set is shown in the title.

reply

Morris

AWDR advises not to put "too much" code into views. But how much is too much?

Suppose I have
class Model < ActiveRecord::Base
belongs_to :a
has_one :b
has_many :c
end

Which is more idiomatically Rails-ish, or more recommended by course staff?

class ModelController < ApplicationController
def action
@model = Model.find(params[:id])
end
end

controllers/model/action.html.erb:
blah blah <% some code involving @model %>
<% @model.a %>
<% @model.b %>
<% @model.cs.each do |c| %>
...
<% end %>

--------- OR ----------
class ModelController < ApplicationController
def action
@model = Model.find(params[:id])
@a = @model.a
@b = @model.b
@cs = @model.cs
end
end

controllers/model/action.html.erb:
blah blah <% some code involving @model %>
<% @a %>
<% @b %>
<% @cs.each do |c| %>
...
<% end %>

---------------------------------------------
Is it okay to expose the model's relationships inside the view, or is it better to hide all of this inside the controller, and have the view just access those related objects as instance variables of their own?

reply

xertroyt

This is a great question, and a hard question.

Three rules of thumb all of which may be ignored in a pinch:

-- Try not to go follow more than one association from an object in a view. I.e., one dot. You can think of this as a version of the Law of Demeter (http://en.wikipedia.org/wiki/Law_Of_Demeter).

-- Iterating through collections is fine.

-- Using if/then/else based on the evaluation of a single variable is fine. Anything more complicated is getting into risk.

Here is a wonderful statement on how you will, in the long run, find it convenient to push code all the way back to the model:

http://weblog.jamisbuck.org/2006/10/18/skinny-c...

reply

Morris

A useful point about partials, which hasn't been mentioned in lecture, and if mentioned in AWDR is not stressed enough, in my opinion:

While copying code from the CRUD example into my assignment 3 code, I found myself putting a lot of stuff into the :locals hash. Then, accidentally, I discovered (and verified experimentally) that any instance variables that are available to the top-level view are also available to the partials. So one doesn't need to use locals to pass those variables to the partial. This also means that when, for instance, using the _section.html.erb partial from the new and edit views, if the instance variable being set in the controller method is @section, and new or edit is calling <%= render partial => :section %>, inside the partial the same information can be accessed as "@section" or as "section", the difference being that if you actually modify that variable, in one case it's local to the partial and in the other, it affects other things (such as if the higher-level view were doing other stuff using that instance variable after it calls render). This is actually how I tripped over this, because I was getting confused over whether I should use the @ or not, and discovered that my output was identical either way.

reply

xertroyt

To my knowledge it is not mentioned in AWDR, though I haven't searched for it.

It's very interesting that you mention that because while I've been using partials for a long time, I never knew that instance variables were even copied over until (guess what?) I was debugging a partial that was surprising me.

I think that as a best practice you want to avoid using instance variables in partials because it creates an invisible dependency. If you use :object and :locals it's a bit more explicit. And, as you point out, you run the risk of changing a variable in the template that is calling the partial.

A nice feature to add to Rails would be a mechanism where the partial could say that it "requires" certain locals and/or instance variables.

reply

Morris

FYI: David Black's "Rails Routing" book, mentioned in this week's lecture, is available on-line (on Safari), FREE through the Harvard library system to anyone with a Harvard ID.

http://nrs.harvard.edu/urn-3:hul.ebookbatch.SAF...

reply

xertroyt

Excellent, thanks for reminding everyone of this.

reply

Student 8

Speaking of routing, any time you change your 'routes.rb' file, *restart your mongrel server.*

Otherwise you may get "Only get, head, post, put, and delete requests are allowed. Ruby on Rails error".

Silly me for not going straight to Google on this one. I kept thinking I had screwed-up the code. I found the solution here: http://tinyurl.com/6a8ff5

reply

Student 8

Still a little lost (like me)? http://www.learningrails.com offers nice step-by-step screencasts on app-building basics.

reply

xertroyt

Those videos look good. Remember, though, that the "last word" is the AWDR book. Also when watching the videos, be cautious about anything that might mention Rails 2.2 (just released). We're on 2.1.1.

reply

hogganbeck

John:

At one point, you used a button in your GUI file window that brought up Terminal already cd'd to that particular directory. What was that? My wrists hurt.

reply

xertroyt

On Windows, you want "Open Command Window Here" from this page: http://www.microsoft.com/windowsxp/downloads/po...

On the Mac, you want "Open Terminal here": http://www.entropy.ch/software/applescript/

reply

hogganbeck

Fantastic. It works! Thx.

reply

Morris

The documentation for methods in ActionView::Helpers subclasses (e.g. FormHelper, DateHelper) shows a first argument for each method, which is an object_name. But in the examples, and in practice, that argument seems to be optional, and defaults to the object which is the first argument of the enclosing form_for method. E.g. the signature for datetime_select is

datetime_select(object_name, method, options = {}, html_options = {})

but in practice we call it as something like

form.datetime_select :observed_at

with the implicit @observation (or :observation) as the object, and :observed_at as the method on that object.

Two questions:

1) How does this work, from a language viewpoint? Do these methods examine the types of their arguments, figure out that the first argument is the name or symbol of a method rather than an instance variable, and adjust the argument list appropriately?

2) More importantly, other than by reading AWDR and intuiting the methods' behavior from the code examples, how would we be able to figure out that some arguments (not the trailing arguments, and not those shown with default values in the method signature via an '=' character) are optional?

reply

xertroyt

1) When you call form.datetime_select, the form already knows what object is in play (because you said form_for, which yields the form variable). So it can delegate directly to the datetime_select which expects an object as the first param. I haven't looked at the source, but that is likely how it is implemented.

2) The on-line documentation is bad, in my humble opinion, because it underplays what the "form builder" (the variable form_for yields). The docs (http://api.rubyonrails.org/classes/ActionView/H...) say:

The generic way to call form_for yields a form builder around a model . . . (and they quote an example) . . .

There, the first argument is a symbol or string with the name of the object the form is about, and also the name of the instance variable the object is stored in.

[KEY SENTENCE -->] The form builder acts as a regular form helper that somehow carries the model. Thus, the idea is that

<%= f.text_field :first_name %>

gets expanded to

<%= text_field :person, :first_name %>

-----

And that is the whole enchilada. You can use any of the form helpers via that "f" or "form" or whatever variable you have assigned to the one yielded to you by form_for.

The RDoc generator that makes the HTML pages should cross-reference what is happening here so that whenever you're looking at "datetime_select" there is some indicator of how you would use it from form_for; and the idea that the variable form_for yields magically sprouts these methods that delegate can't be overplayed enough.

reply

Morris

Even worse, in the on-line doc:
I just inserted a line in one of my views, inside the form_for block, that says

f is a <%= f.type %>
(where f is my block parameter), and it tells me, "f is a ActionView::Helpers::FormBuilder", but api.rubyonrails.com doesn't have ANY DOCUMENTATION for the FormBuilder class. So your idea of cross-referencing FormHelper#text_field, etc., is a good idea, but there's NOTHING TO CROSS-REFERENCE.

(and, to add to the list of complaints: instead of just ranting about this here, I went over to rubyonrails.com to report this gap in the documentation, like a good citizen, and I can't find any easy way to report documentation bugs. I'm not feeling sufficiently motivated to create an account and log in, just to inject a documentation complaint into the on-line Rails bug-tracking system.)

reply

Morris

After watching this week's lecture this morning, and checking quickly in the Rails documentation, I wonder why there's a helper for Atom feeds, but no helper for RSS. It seems to me that I see lots more RSS feeds on the net than Atom feeds.

A very quick web search reveals this article

http://myersds.com/notebook/2006/05/11/how_to_g...

which seems very clear, which is just using generic Builder to put together the XML for an RSS document, which raises the next question, whether AtomFeedBuilder is providing any real added value, compared to just building an Atom feed out of raw XML

reply

xertroyt

Atom is a building block for providing much more sophisticated feeds.

This is a biased account, but will give you some of the flavor for why you might go for Atom:

http://www.intertwingly.net/wiki/pie/Rss20AndAt...

This is a big deal: http://www.intertwingly.net/wiki/pie/Rss20AndAt...

Re: building an Atom feed out of raw XML: If you use the atom_feed helper, then if common usage moves to a different version of Atom, you'll have built your feed using a higher level API, which might get updated -- and then you don't have to dig around in your own home-build XML.

There is nothing wrong with providing both kinds of feeds. "One resource, many representations." If you include two "auto discovery" links, must feed consumers will give you an option to select (atom vs rss 2.0, etc.).

reply

Morris

Which still leaves the question of why the developer(s?) of Rails included a helper for Atom in the Rails core classes, but chose not to do the same for RSS. The same issue, of common usage moving to a different version, applies.

reply

xertroyt

Well, one reason might be that there isn't much of a standard for RSS: http://en.wikipedia.org/wiki/Rss#Variants

If you want to support RSS, which one? Should you expose 0.9 AND 1.0 AND 2.0? Do we need to support ALL of those? There is doubtless a plugin that will do all of that for you; I would bet that DHH would say that Rails, being "opinionated software" advocates Atom . . .

reply

philadelphia

Hi there,

I'm thinking about the final project, and as one of the elements, I'm interested in maintaining minute by minute usage data associated with internet enabled equipment. But I'd like to roll the data up into hours, days, weeks, and years, so that I will not have to maintain hundreds of thousands of rows for each user for each year.

Is there a best practices way of doing this? I've been thinking about writing it myself using after_save callbacks on a Usage table. But, it's not all together trivial due, amongst other things, to some recursion issues I foresee (I foresee, I foresee, I foresee :-). I'd be very happy to find out someone has already figured this out and has written a gem or plugin...

Thanks,
Mike

reply

xertroyt

What you want to do is collect all of the observations, and then periodically have a process that summarizes the data. Or, if you wanted to do it on the fly -- that's fine, too. But i think you are going to have to hand-roll this process. This is very much like data-warehousing ETL (extract / transform / load).

The original collection could be into a flat file or into a database table (you would truncate or archive the original data after you summarize it). You would surely want to archive it.

The standard way to manage rollups is with a table such as the one used in MetricsMine. You are free to use it. Note that that table doesn't have indices. You would want to keep your data in a database such as Postgresql or MySQL which does allow for indexes.

In any case, the big question for you will be what you want to do with the immediately-collected data before it gets summarized. Will users want to see their data trickle in within the hour?

reply

philadelphia

They will want to see minute by minute data for the previous 60 minutes. Then hourly averages for the previous day. The daily for the previous 7 days. Then weekly (forever?).

The difficulty is the missing data... machines go down and so do internet connections. Therefore, the data will have missing elements, which in some cases will span hours, days or weeks. So, I need to be able to handle missing observations as well.

I'll take another look at the rollups in MetricsMine. I may be able to use some or all of it for this purpose.

Thanks,
Mike

reply

philadelphia

After reading your comment and thinking about it some more, maybe the best way to handle this is to write to the DB for each event with a timestamp and then report based upon this information. Every week a cron job would run and create weekly average. In that way, the max rows in the table would be 10,800 per user per reporting device. That's not too bad, as long as they are indexed. Most of the reporting will pull the most recent record (greatest timestamp) and will not require a lot of overhead (if it is indexed).

Thanks,
Mike

reply

xertroyt

That sounds good to me. I think a real question is going to be granularity of the initial data, and the smallest size of the initial report. You would have a significant boost if you could make that initial report by even 10-minute intervals.

reply

xertroyt