Rails sucks, NodeJS rules?

[Merlyn Albery-Speyer]

I've implemented Jobagator.org in both Rails and NodeJS now. The NodeJS version is sooo much faster: http://www.curiousattemptbunny.com/2012/10/jobagator-in-nodejs-is-notably-faster.html

[Matthew Boeh]

You know, I've read this email, and I've read your post, and I've
looked at the screenshots very carefully, and I just have to say:

Is this a joke?

Because I really hope it is.

> --
> You received this message because you are subscribed to the Google Groups
> "Portland JavaScript Admirers" group.
> To post to this group, send email to pd...@googlegroups.com.
> To unsubscribe from this group, send email to
> pdxjs+un...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

[Sidu Ponnappa]

> Is this a joke?
Everyone knows that Rails doesn't scale.

Best,
Sidu.
http://sidu.in
http://twitter.com/ponnappa

> You received this message because you are subscribed to the Google Groups "pdxruby" group.
> To post to this group, send email to pdx...@googlegroups.com.
> To unsubscribe from this group, send email to pdxruby+u...@googlegroups.com.

[avitevet]

What's taking time on the server?

In my experience it can be difficult to narrow down the cause of the server side slowness... I reduced view render time on my dev machine from 5+ seconds to 1.3s on this page (http://fitlogr.com/leaderboard/22) by converting a frequently called partial into a helper function.  It was an unintuitive action to take based on the perf testing tools and the server log, because they consistently showed that even though there were dozens of renders of this partial, each rendering took only 5ms (< 1 sec total for all renders).  I guess it was the rails overhead in calling a partial?

Avi

[Matthew Boeh]

I might as well chip in here: aside from the many, many issues I have
with the method and presentation of this comparison, what exactly is
causing 6KB of HTML encompassing 10 records to take upwards of 3
seconds to render and transfer?

I've written (and fixed) very, very slow views in my time and I still
can't even comprehend what would make that page take any more than
~100-200ms to render, unless you're making external HTTP calls in
process or something.

> --
> You received this message because you are subscribed to the Google Groups
> "pdxruby" group.
> To post to this group, send email to pdx...@googlegroups.com.
> To unsubscribe from this group, send email to
> pdxruby+u...@googlegroups.com.

> To view this discussion on the web visit
> https://groups.google.com/d/msg/pdxruby/-/BIwensfOgiYJ.

[Paul Jungwirth]

> even though there were dozens of
> renders of this partial, each rendering took only 5ms (< 1 sec total for all
> renders). I guess it was the rails overhead in calling a partial?

One thing I've seen before with partials called many times in one
render is that they are all extremely fast except for one, which is in
the hundreds of ms range. I eventually figured out that Ruby was
running garbage collection, and that's where the hit was coming from.
I'm not sure what happens under the hood when a partial gets rendered,
but it seems likely a helper function is lower overhead.

Paul

> On Thursday, October 11, 2012 9:28:57 PM UTC-7, Merlyn Albery-Speyer wrote:
>>
>> I've implemented Jobagator.org in both Rails and NodeJS now. The NodeJS
>> version is sooo much faster:
>> http://www.curiousattemptbunny.com/2012/10/jobagator-in-nodejs-is-notably-faster.html
>

> --
> You received this message because you are subscribed to the Google Groups
> "pdxruby" group.
> To post to this group, send email to pdx...@googlegroups.com.
> To unsubscribe from this group, send email to
> pdxruby+u...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/pdxruby/-/BIwensfOgiYJ.
>
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

--
_________________________________
Pulchritudo splendor veritatis.

[markus]

M --

> Is this a joke?
>
> Because I really hope it is.

Joke may be too narrow of a term. Koan might be closer. At least,
having been acquainted with Merlyn for some years I find myself
categorizing him similarly to how I expect people to categorize me: a
source of provocative utterances that you shouldn't accept or reject
without thinking about, with the expectation that more often than not
the act of thinking will turn out to have been the payload and the
original provocative utterance was mostly just a delivery mechanism.

-- Markus

[Matthew Boeh]

Well, that certainly makes me feel better. It did make me think,
that's for sure.

I am weary of "X is faster than Y, these numbers prove it" material.
One might as easily talk about Ruby being shinier than Java, or Python
being more purple than a python. Spurious benchmark posts show up on
blogs, then on Hacker News, then on Twitter, and they're difficult to
avoid and drown out meaningful discussion. I really wasn't happy to
see that kind of thing showing up here.

My question was intended literally. The comparison was so bizarre and
presented in such a provocative way that I couldn't possibly imagine
it being serious.

> You received this message because you are subscribed to the Google Groups "pdxruby" group.
> To post to this group, send email to pdx...@googlegroups.com.
> To unsubscribe from this group, send email to pdxruby+u...@googlegroups.com.

[Bryan Donovan]

Compare the Node version to the exact same API/Backbone-based setup with Sinatra-Synchrony or Goliath, and then you might be comparing apples to apples.

On Thursday, October 11, 2012 9:28:57 PM UTC-7, Merlyn Albery-Speyer wrote:

[markus]

> My question was intended literally. The comparison was so bizarre and
> presented in such a provocative way that I couldn't possibly imagine
> it being serious.

I meant to point this out in my first response, and exhibit A:

> >> > http://www.curiousattemptbunny.com/2012/10/jobagator-in-nodejs-is-notably-faster.html


People who choose nicks like "curiousattemptbunny" are likely more
inclined toward engaging in conceptual exploration and provoking thought
than being taken as a authoritative, "I have told you the truth, you may
stop thinking now" source.

So I think he is serious, but (and Merlyn, feel free to jump in here and
correct me if I'm wrong) what he's serious about isn't "OMG yous must
all stop rails now, go node for greater good ok thx bye" but rather
"Here is an observed fact about the world with all the messy bits still
attached; not a controlled experiment but rather an observation made in
passing about something that really happened. It is somewhat surprising
and thus bears thinking about."

-- Markus

P.S. On the original puzzle, I think the key insight is to ignore the
node.js comparison as a red herring (since the surprise is how slow the
rails version is) and ask what could CompaniesControler#index be doing
that takes so darned long? If it's just bad code explicitly doing
needless work, shame on you Merlyn, but if it's some "favour" that
active record is doing for us or some other pit trap, that might be
interesting to know about. So, Merlyn, what's the index method doing?

[Sean McCleary]

Is this some kind of slow-off that compares the speed of two completely different things? This article must be troll bait. Reasonable people do not say that technology stack A is faster than a completely refactored different set of technology stack B. The better question might be how do you possibly get Rails to take 4.5 seconds to render a 5.29kb page? Do you run your rails app on a TI-83? Now that would be an interesting article indeed.

- Sean McCleary

--

You received this message because you are subscribed to the Google Groups "pdxruby" group.
To post to this group, send email to pdx...@googlegroups.com.
To unsubscribe from this group, send email to pdxruby+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msg/pdxruby/-/DKENdm6McsQJ.

[Jesse Cooke]

You know, we rarely hack at Ruby meetings. Maybe at the next one we dive into this?

[Igal Koshevoy]

On Thu, Oct 11, 2012 at 9:28 PM, Merlyn Albery-Speyer <curious.at...@gmail.com> wrote:

I've implemented Jobagator.org in both Rails and NodeJS now. The NodeJS version is sooo much faster: http://www.curiousattemptbunny.com/2012/10/jobagator-in-nodejs-is-notably-faster.html

To Merlyn and others that are considering writing "X sucks, Y rules" posts: please don't, this isn't constructive.

I and others would be better able to help and more eager to assist if you post help requests with content like:

My Rails app is taking over three seconds to render a trivial index page with a list of 10 items. This seems wrong. Can someone suggest some tools and techniques to speed up my app or identify the problem? Attached is the section of the Rails log and NewRelic output showing which part of the app they think is slow, and the code these claim is responsible for the slowdown. Can someone that knows Rails please meet up with me to look over my code at one of the two monthly Ruby meetups, or three weekly hackathons? Would someone that knows how to fix these sorts of problems be interested in a quick contract job to help me out? Thanks!

Anyway, the ball is in Merlyn's court and I'm unsure that it makes much sense for others to keep responding to this thread until he provides more information.

Thanks,

-igal

[markus]

> The better question might be how do you possibly get Rails to take 4.5
> seconds to render a 5.29kb page?

Yeah. I think the node stuff is a red herring. 4.5 seconds to render a
smallish page is along the lines of taking seven months to make a
sandwich--there has to be more to it than just a tricky lid on the
mustard jar. My big question is, is there something that would be
"obvious" from the source of CompaniesController#index, or is it
something being done by one of the underlying magic layers?

-- Markus

[Matthew Boeh]

Conceptual exploration is meaningless if it doesn't start from a
position of good faith and at least a little bit of meaningful data.
In this case:

* We are comparing two solitary page loads against unknown apps
running on unknown stacks.
* The app designs are clearly different. It's not useful to say "I
know this is apples and oranges" when you proceed to make a graphic
saying "look, this orange is so much more bumpy than the apple".
* The numbers were collected with a primed cache, as evidenced by the
many 304 responses. I could assume that neither app has any meaningful
optimization there, but I don't have any reason to.
* The Rails app's total includes requests to external analytics
services like Google Analytics and Mixpanel, while the Node app's does
not. This is mentioned nowhere, but the "total" numbers are
highlighted as if they mean something.

You're right that the Rails/Node aspect of the comparison is a red
herring. I'd go a little further and call it specious. I would have
been glad to hear about how Merlyn rewrote his app to make it much
faster. That would have been an interesting conversation, and one I'd
still be glad to have.

In this case, we have an opaque, misleading ball of data represented
in a manner obviously likely to make people upset and angry. The
question mark doesn't make "Rails sucks, NodeJS rules" any less
confrontational. It was cross-posted to both the JS and Ruby mailing
lists. As a longtime troll, I recognize trolling when I see it.

Igal:

Thank you, and I'll let the matter rest here.

[Ben Munat]

+1... though I never seem to make it to meetings anymore.

b

On 10/12/12 10:16 AM, Jesse Cooke wrote:
> You know, we rarely hack at Ruby meetings. Maybe at the next one we dive
> into this?
>
>
> On Fri, Oct 12, 2012 at 10:04 AM, Sean McCleary <sea...@gmail.com
> <mailto:sea...@gmail.com>> wrote:
>
> Is this some kind of slow-off that compares the speed of two
> completely different things? This article must be troll bait.
> Reasonable people do not say that technology stack A is faster than
> a completely refactored different set of technology stack B. The
> better question might be how do you possibly get Rails to take 4.5
> seconds to render a 5.29kb page? Do you run your rails app on a
> TI-83? Now that would be an interesting article indeed.
>
> - Sean McCleary
>
>
>
> On Fri, Oct 12, 2012 at 9:35 AM, Bryan Donovan <bdo...@gmail.com
> <mailto:bdo...@gmail.com>> wrote:
>
> Compare the Node version to the exact same API/Backbone-based
> setup with Sinatra-Synchrony or Goliath, and then you might be
> comparing apples to apples.
>
>
> On Thursday, October 11, 2012 9:28:57 PM UTC-7, Merlyn
> Albery-Speyer wrote:
>
> I've implemented Jobagator.org in both Rails and NodeJS now.
> The NodeJS version is sooo much faster:

> http://www.__curiousattemptbunny.com/2012/__10/jobagator-in-nodejs-is-__notably-faster.html

> <http://www.curiousattemptbunny.com/2012/10/jobagator-in-nodejs-is-notably-faster.html>
>
>
> --
> You received this message because you are subscribed to the
> Google Groups "pdxruby" group.
> To post to this group, send email to pdx...@googlegroups.com

> <mailto:pdx...@googlegroups.com>.

> To unsubscribe from this group, send email to
> pdxruby+u...@googlegroups.com

> <mailto:pdxruby%2Bunsu...@googlegroups.com>.

> To view this discussion on the web visit
> https://groups.google.com/d/msg/pdxruby/-/DKENdm6McsQJ.
>
> For more options, visit https://groups.google.com/groups/opt_out.
>
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "pdxruby" group.
> To post to this group, send email to pdx...@googlegroups.com

> <mailto:pdx...@googlegroups.com>.

> To unsubscribe from this group, send email to
> pdxruby+u...@googlegroups.com

> <mailto:pdxruby%2Bunsu...@googlegroups.com>.

[Igal Koshevoy]

On Fri, Oct 12, 2012 at 10:16 AM, Jesse Cooke <je...@jc00ke.com> wrote:

You know, we rarely hack at Ruby meetings. Maybe at the next one we dive into this?

There's a Ruby Beginners Meetup next Tuesday <http://calagator.org/events/1250462852>. It would be great if Merlyn (CC'ed) could drop by for us to help debug the Rails performance issue. And then have someone present how they debugged and fixed the problem at the next general Ruby meeting.

Collaborative hacking as part of a meeting can be awesome, but seems to only work well when the group is small because that produces a high level of engagement, whereas a large group seems to not work well because it's mostly one person coding, a few people heckling, and a lot of bored, quiet people. With 50+ people attending the general meetings, I'm hesitant to add a potentially lengthy and unbounded hacking activity that may not be of interest to many as part of the main agenda. That said, I'd be willing to try doing a hacking activity at the very end of the meeting as a "optional" activity after all the presentations are done for the evening, so that only those interested in the particular task stay.

Thoughts?

-igal

[Jesse Cooke]

Regardless of the way Merlyn presented it, I still think we could all learn a lot from dissecting this.

--

[Igal Koshevoy]

On Thu, Oct 11, 2012 at 11:12 PM, Matt Youell <ma...@youell.com> wrote:

If you haven't already, you should try New Relic. Collects these stats from inside your app so it's more consistent and you can experiment with changes and see what worked pretty quickly. (I have nothing to do with them, it's just an awesome tool.)

It may be time for our annual presentation on, "Why is my Ruby app slow? What is this NewRelic thing and how do I use it? What are reasonable and effective ways to speed things up?" 

We have many people that are either new to the group or Ruby, haven't heard this kind of talk before, and would find it useful. And for those that have heard it before, it'd be great if we could also get an update on advances made in this field, especially since the last time we heard this talk, Ruby 1.9 and Rails 3 were still bleeding edge, but introduced a lot of enhancements to make more sophisticated performance monitoring possible.

Who's interested in presenting? Last year's rendition by NewRelic's team was great.

We've got up to an hour of time available on the November meeting agenda, and plenty more on the December agenda.

-igal

[Jesse Cooke]

On Fri, Oct 12, 2012 at 10:24 AM, Matthew Boeh <m...@mboeh.com> wrote:

Conceptual exploration is meaningless if it doesn't start from a
position of good faith and at least a little bit of meaningful data.
In this case:

* We are comparing two solitary page loads against unknown apps
running on unknown stacks.
* The app designs are clearly different. It's not useful to say "I
know this is apples and oranges" when you proceed to make a graphic
saying "look, this orange is so much more bumpy than the apple".
* The numbers were collected with a primed cache, as evidenced by the
many 304 responses. I could assume that neither app has any meaningful
optimization there, but I don't have any reason to.

This: Good catch 

* The Rails app's total includes requests to external analytics
services like Google Analytics and Mixpanel, while the Node app's does
not. This is mentioned nowhere, but the "total" numbers are
highlighted as if they mean something.

This: Extremely important. I usually disable external javascripts in development

since they slow down page load so much.

In fact, do we know what environment the Rails app was running in? 

You're right that the Rails/Node aspect of the comparison is a red
herring. I'd go a little further and call it specious. I would have
been glad to hear about how Merlyn rewrote his app to make it much
faster. That would have been an interesting conversation, and one I'd
still be glad to have.

In this case, we have an opaque, misleading ball of data represented
in a manner obviously likely to make people upset and angry. The
question mark doesn't make "Rails sucks, NodeJS rules" any less
confrontational. It was cross-posted to both the JS and Ruby mailing
lists. As a longtime troll, I recognize trolling when I see it.

Even though it's a troll, there are still things to learn. They don't need to

all be epiphanies. 

[markus]

> We've got up to an hour of time available on the November meeting
> agenda, and plenty more on the December agenda.

I would like to take ten minutes or so to draw some crude pictures of
islands and sailboats and stuff (and tell a little story about how
clever people can be) which might make a good lead in to the topic if
anyone is interested.

-- Markus

[Jason LaPier]

On Fri, Oct 12, 2012 at 10:49 AM, Igal Koshevoy <ig...@pragmaticraft.com> wrote:

On Fri, Oct 12, 2012 at 10:16 AM, Jesse Cooke <je...@jc00ke.com> wrote:

You know, we rarely hack at Ruby meetings. Maybe at the next one we dive into this?

There's a Ruby Beginners Meetup next Tuesday <http://calagator.org/events/1250462852>. It would be great if Merlyn (CC'ed) could drop by for us to help debug the Rails performance issue. And then have someone present how they debugged and fixed the problem at the next general Ruby meeting.

That's a great idea, I hope Merlyn agrees to it. The original post was a little like opening a mystery novel where someone is murdered on page one, and then the rest of book is about how a new security system got installed and no one was ever killed again. That might be a great story for the pdxjs list, but for the pdxruby list I think we're all going, "Wait, who killed CompaniesController#index?!?"

[Jesse Cooke]

On Fri, Oct 12, 2012 at 10:49 AM, Igal Koshevoy <ig...@pragmaticraft.com> wrote:

I was thinking that, instead of doing presentations at all, we just hack for a session. We could do smaller groups of 3-5 people, pick a project or several, write some documentation, fix some bugs. It might be nice to try something different once in a while.

-igal

[Jesse Cooke]

On Fri, Oct 12, 2012 at 11:32 AM, markus <mar...@reality.com> wrote:

> We've got up to an hour of time available on the November meeting
> agenda, and plenty more on the December agenda.

I would like to take ten minutes or so to draw some crude pictures of

*Only* 10 minutes? ;) 

islands and sailboats and stuff (and tell a little story about how
clever people can be) which might make a good lead in to the topic if
anyone is interested.

-- Markus

[Igal Koshevoy]

On Fri, Oct 12, 2012 at 11:32 AM, markus <mar...@reality.com> wrote:

Yes, that'd be great.

I'd also enjoy a "Here's my simple Ruby program that has a performance problem, and I've obscured a few bits of code that I'm sure aren't important *wink*, but I suppose you could try to guess what they are" talk with you twirling your mustache.

-igal

PS: Does one of the people on the boats get scurvy and is his name Matt? Would you be willing to give that talk again at pdxruby some time? It was awesome.

[Sam Livingston-Gray]

On Fri, Oct 12, 2012 at 8:13 AM, Igal Koshevoy <ig...@pragmaticraft.com> wrote:
> On Thu, Oct 11, 2012 at 11:12 PM, Matt Youell <ma...@youell.com> wrote:
>>
>> If you haven't already, you should try New Relic. Collects these stats
>> from inside your app so it's more consistent and you can experiment with
>> changes and see what worked pretty quickly. (I have nothing to do with them,
>> it's just an awesome tool.)

I'm not seeing Matt's response in the thread above this, so this may
have been taken out of context, but: last I checked, Merlyn works at
New Relic. ;>

> It may be time for our annual presentation on, "Why is my Ruby app slow?
> What is this NewRelic thing and how do I use it? What are reasonable and
> effective ways to speed things up?"

+1, especially since New Relic has so many developers in town.

[Igal Koshevoy]

On Fri, Oct 12, 2012 at 12:36 PM, Sam Livingston-Gray <gee...@gmail.com> wrote:

On Fri, Oct 12, 2012 at 8:13 AM, Igal Koshevoy <ig...@pragmaticraft.com> wrote:
> On Thu, Oct 11, 2012 at 11:12 PM, Matt Youell <ma...@youell.com> wrote:
>>
>> If you haven't already, you should try New Relic. Collects these stats
>> from inside your app so it's more consistent and you can experiment with
>> changes and see what worked pretty quickly. (I have nothing to do with them,
>> it's just an awesome tool.)

I'm not seeing Matt's response in the thread above this, so this may
have been taken out of context

Oops. Looks like that was sent just to pdxjs, but not pdxruby. Matt was directly responding to the original message, so no additional context was provided.

 

but:  last I checked, Merlyn works at New Relic.  ;>

Huh, Linkedin says he is. Hmmm. My leg, it feels like it's being pulled on.

-igal

[Brian Artiaco]

TL;DR

Don't feed the trolls.  Or perhaps merlyn's wit is too dry to be properly appreciated via text. :)

[markus]

> > but: last I checked, Merlyn works at New Relic. ;>

> Huh, Linkedin says he is. Hmmm. My leg, it feels like it's being
> pulled on.

Really? Do you suppose curiousattemptbunny might be trying to lure us
down the rabbit hole? I'm shocked, shocked I tell you! :)

Hmmm. I wonder if there's anything worth knowing about lurking down
there?

-- Markus

[Matthew Boeh]

I'd sure like to know, but I lack the ability to investigate
performance issues with code that I can't see or run for myself.

I'll admit that may be a failing.

[ben]

I don't know what Merlyn is up to, but I for one would like to congratulate Markus on a fantastic combo thread-hijack/trolling.

[Igal Koshevoy]

On Fri, Oct 12, 2012 at 1:24 PM, markus <mar...@reality.com> wrote:

> > but:  last I checked, Merlyn works at New Relic.  ;>
> Huh, Linkedin says he is. Hmmm. My leg, it feels like it's being
> pulled on.

Really?  Do you suppose curiousattemptbunny might be trying to lure us
down the rabbit hole?  I'm shocked, shocked I tell you!  :)

I want pdxruby to be able to help people with legitimately Ruby issues. Anything that makes people afraid to ask legitimate Ruby questions, or prevents people from trying to help others in good faith here is annoying. Unfortunately, many people use the "X sucks" format because they don't know better or are following others as a bad example, which is why my only response on that matter was written in hopes of providing guidance to them on asking better questions. As for Mr. curiousattemptbunny, he knows better than to do this. Meh.

Hmmm.  I wonder if there's anything worth knowing about lurking down there?

So far all I've found while rummaging around down there was a "Rails doesn't scale" quip with a 2007 timestamp. It smells stale.

A recent, similarly-themed thread on 4chan was much classier because the attached code sample included an attempt to "rm -rf /", but had a typo, which people then submitted fixes to, and then offered a cross-platform implementation using FileUtils.rm_rf, and then began arguing about whether the typo and lack of cross-platform support in the Trojan was itself a troll. But this thread here? It could use more sage.

-igal

[markus]

> But this thread here? It could use more sage.

I tried to add sage, but I ran out of thyme.

-- Markus

P.S. I've also been trying to throw in a dash of Lighten Up but I fear
the day is too grey for some of us to stomach anything but Serious with
a dollop of Glum. Ah, Portland sunshine, how we miss you. :(

[Ben Munat]

Ouch.

[Robb Shecter]

Ugh. Just noticed that the next meeting (Tues., Nov. 6) is election day.

I was thinking about presenting on this, but I'll be at an election party.

[Igal Koshevoy]

On Fri, Oct 12, 2012 at 5:00 PM, Robb Shecter <ro...@weblaws.org> wrote:

Ugh. Just noticed that the next meeting (Tues., Nov. 6) is election day.

I'm going out on a limb here, but I'm sure that someone will get elected, regardless of what we do. So come by and talk Ruby.

I was thinking about presenting on this, but I'll be at an election party.

Oh fine. How about presenting in December? 

-igal 

[Matthew Boeh]

This is an issue for me as well.

On Fri, Oct 12, 2012 at 5:00 PM, Robb Shecter <ro...@weblaws.org> wrote:

> Ugh. Just noticed that the next meeting (Tues., Nov. 6) is election day.
>
> I was thinking about presenting on this, but I'll be at an election party.
>
>

> --
> You received this message because you are subscribed to the Google Groups
> "pdxruby" group.
> To post to this group, send email to pdx...@googlegroups.com.
> To unsubscribe from this group, send email to
> pdxruby+u...@googlegroups.com.

> To view this discussion on the web visit

> https://groups.google.com/d/msg/pdxruby/-/X7I1b-L6scMJ.

[Igal Koshevoy]

Matt,

Rats. I have you down for doing a Celluloid talk that night. Should I reschedule your talk to December?

-igal

[Matthew Boeh]

Yes, please. I hadn't realized that the first Tuesday in November was,
well, the first Tuesday in November.

[Igal Koshevoy]

Matthew,

I've moved your Celluloid talk to the December 4th meeting.

Thanks!

-igal

[Robb Shecter]

It's Portland; we're _obligated_ to go to election events.

But yes, Igal, I'm down for doing a presentation in December if you still need it. (I see that Matt's been moved to Dec.) I was offering to do one on the topic of optimizing Rails apps, but could do others as well.

I'd call myself a sort-of-expert on optimization: I can show all the levels that need to be considered and how I handle them in a 400,000 page Rails app, from DNS TTLs to database I/O. I've also been a customer of New Relic for a couple of years and can talk about the advantages and disadvantages of their service.

[markus]

> I was thinking about presenting on this, but I'll be at an
> election party.

> This is an issue for me as well.

If everyone else bails, I suppose I could draw sailboats and barges and
stuff for a little longer. I might even be able to make up some numbers
to support my thesis. :)

-- Markus

[Matt Youell]

On 10/12/12 12:36 PM, Sam Livingston-Gray wrote:
>
> I'm not seeing Matt's response in the thread above this, so this may
> have been taken out of context, but: last I checked, Merlyn works at
> New Relic. ;>

Well, if nothing else I have the pleasure of knowing I've given Merlyn a
chuckle. :)


--
-/matt/-
http://youell.com/matt

[john melesky]

On Fri, Oct 12, 2012 at 05:14:56PM -0700, Matthew Boeh wrote:
> Yes, please. I hadn't realized that the first Tuesday in November
> was, well, the first Tuesday in November.

*Technically*, Election Day in the US is held on the Tuesday
immediately following the first Monday of November, due to legal
technicalities surrounding Electoral College law (or
somethingorother).

Thank you for your attention.

-john