Disqus - Latest Comments for tav

Re: Ruby-style Blocks in Python

tav — Sat, 17 Jan 2015 18:02:20 -0000

Nice work!

Re: You are dangerously bad at cryptography | Happy Bear Software | Web Application Development

tav — Mon, 27 May 2013 20:49:00 -0000

This article provides some relevant background info and links to some of the papers: http://rdist.root.org/2010/...

And here is the talk/slides from the talk at Blackhat 2010: http://www.youtube.com/watc...

In other words, timing attacks over the internet are surprisingly feasible.

Re: RIP, Aaron Swartz

tav — Sat, 12 Jan 2013 08:23:12 -0000

We've lost a bright light. Thank you Cory for your beautiful and heartfelt words.

Re: OAuth 3.0: The Sane and Simple Way To Do It

tav — Sat, 28 Jul 2012 20:28:49 -0000

This article was last modified on July 11th 2010 according to https://github.com/tav/blog...

The presence of a date/timestamp in the file on GitHub is exactly what my blogging tool uses to determine whether an article is a draft or published, i.e. shown on the frontage, rss feed, archive listing, etc.

Unfortunately, I had committed this as a draft in order to get some early feedback back then and thus why it's still publicly visible.

Re: OAuth 3.0: The Sane and Simple Way To Do It

tav — Sat, 28 Jul 2012 17:33:34 -0000

In the years since I drafted this, companies like Google have started providing Discovery APIs for their services, e.g. https://developers.google.c...

If I were rewriting this draft today, I'd probably incorporate some elements from that in an effort to standardise writing API clients.

Re: Crawling, Culling, and Machine Learning in JavaScript

tav — Mon, 26 Sep 2011 13:06:34 -0000

An article going into the details of the filtering/recommendation system — including the various alternative routes you tried and why/what didn't work — would perhaps be interesting...

Re: Why Bitcoin Will Fail As A Currency

tav — Tue, 07 Jun 2011 11:05:22 -0000

Well, sort of, except when you think about it. Almost a third of total possible Bitcoins has already been mined. The individuals who own these Bitcoins could clearly afford computers, which are still today more valuable than the Bitcoins.

Thus, it makes sense that the significant majority of these individuals shouldn't really need to use their Bitcoins in order to purchase anything right now. They should be able to hold onto those Bitcoins until they are comfortable with the return on investment...

Re: Why Bitcoin Will Fail As A Currency

tav — Tue, 07 Jun 2011 10:54:13 -0000

For some reason a large number of people seem to think that this aspect is important. I really don't understand why, but anyway, I hope this is clear:

* It doesn't matter if the total number of possible Bitcoin units is 21 million or 21 sextillion. If I currently have 1% of the total Bitcoins that can ever be in circulation and decide to hold onto it, I will continue to have 1% whether you divide the units up to 8 or a million decimal places.

That is, the units don't matter, it is the proportion that is being hoarded which does.

Re: Why Bitcoin Will Fail As A Currency

tav — Tue, 07 Jun 2011 10:12:44 -0000

Those who hoard and those who discard are not necessarily in the same set. So it is perfectly rational to expect — in fact we know this to be true.

Re: Why Bitcoin Will Fail As A Currency

tav — Tue, 07 Jun 2011 10:10:06 -0000

Well, Bitcoins would be backed by computational power if I could use it to access computational power. Similar to say how the Terra Reference Currency guarantees that I would be able to access resources like oil in exchange for it. So I fail to see how Bitcoin is backed by computational power.

Next up, whilst it is sort of related. I am not talking about the deflationary spiral. Also, I almost daily buy things that I know will get cheaper. This computer that I am typing on would be one example...

Re: Why Bitcoin Will Fail As A Currency

tav — Tue, 07 Jun 2011 10:01:20 -0000

There is a massive distinction between knowing with a 100% certainty what share you have of the total Bitcoin and having a relative guess of the share of natural resources that you have — especially since the guess may be made invalid at any point. Sure the Earth is finite, but finding new oil fields in the Arctic certainly changed the market.

Perhaps we will be able to mine a passing asteroid for gold and discover that it has more gold on it than we've extracted so far in human history. Who knows. The point is simply that the lack of certainty encourages more economic activity. Whereas certainty encourages you to hoard.

Re: Promises - Nick's Blog

tav — Wed, 18 May 2011 04:15:45 -0000

Yay! Welcome back!

Re: Fabric Python with Cleaner API and Parallel Deployment

tav — Thu, 24 Feb 2011 18:24:24 -0000

Thanks for pointing out func, Carwyn, I didn't know about it before!

Re: Fabric Python with Cleaner API and Parallel Deployment

tav — Thu, 24 Feb 2011 18:20:28 -0000

Grig, that should be the right output, no? For me, running both fab and fab --list without any fabfiles, results in:

Fatal error: Couldn't find any fabfiles!

Aborting.

Does fab --list output something different for you? Or, were you expecting a different behaviour?

Re: Fabric Python with Cleaner API and Parallel Deployment

tav — Thu, 24 Feb 2011 00:32:32 -0000

Actually, the use of optcomplete within Fabric doesn't make use of any of my extensions to optcomplete (which I use in other projects) — so Fabric would work fine with the version on pypi. Woo!!

Re: Fabric Python with Cleaner API and Parallel Deployment

tav — Thu, 24 Feb 2011 00:29:16 -0000

Hey Scott,

Sorry for the inconvenience caused. See my response to Grig above and do let me know if it works for you. Thanks!

Re: Fabric Python with Cleaner API and Parallel Deployment

tav — Thu, 24 Feb 2011 00:24:24 -0000

Hey Grig,

I've now updated the setup.py in pylibs to also compile the included PyCrypto package. I've added the extra step of running python setup.py to the above doc too. Do let me know if that works for you. Cheers!

Re: Fabric Python with Cleaner API and Parallel Deployment

tav — Tue, 22 Feb 2011 14:14:28 -0000

Hey man,

Nice work. Also, see my comment to Brian K. Jones. Thanks again!

Re: Fabric Python with Cleaner API and Parallel Deployment

tav — Tue, 22 Feb 2011 14:13:54 -0000

The reason that I hadn't done this is because I also make use of personal forks of other packages, e.g. optcomplete, which are also contained within the pylibs repo. Perhaps a proper Fabric fork which refers to the third party dependencies on GitHub via setup.py would solve that issue though.

Or, maybe just let traviscline do all the work =)

Re: Fabric Python with Cleaner API and Parallel Deployment

tav — Tue, 22 Feb 2011 14:09:56 -0000

I actually developed all this to support an Heroku/AppEngine-like single command deployment framework. I was thinking of possibly writing a follow-up once I'd finished developing that to step people through the various benefits. Would you/others be interested in such an article?

Re: Why Google can’t build Instagram

tav — Fri, 12 Nov 2010 04:50:23 -0000

Hey Robert, there's a typo in the link to Instagram — should be http://instagr.am/

Re: Under the hood with App Engine APIs - Nick's Blog

tav — Sat, 25 Sep 2010 03:08:42 -0000

Nice article as always. I've abused the API interface in the past to do things like parallel datastore queries and it's nice to know that you think the interface will be relatively stable going forward... thanks!

Minor typo: `MakeAsyncCall` should be `MakeSyncCall`.

Re: Will You Peerfund My Work?

tav — Sat, 24 Jul 2010 03:14:53 -0000

Hey Nick,

Welcome back!

The money will be used to cover rent, food and utility bills for myself and a few other members of the core team for the next 3 months.

Usually I'd go and do freelance gig(s) for a few months and use the surplus money to support further months of development. But that usually results in dissipation of energy/focus. And I'm quite pleased with the progress being made during this particular iteration and want to maintain it to a state of completion.

So, failing to get peerfunding, I'll be trying the banks again to see if they'll increase their lines of credit. And, failing that, rustle up some freelance gig and try very very hard to keep the collective momentum alive.

Hope that answers your question?

I also put together a document yesterday which hopefully summarises what I'm trying to do in just six slides:

* http://tav.espians.com/peer...

If you could let me know what you think of that document, that'd be super awesome. Thanks!

-- Hope you're well, love, tav

Re: Will You Peerfund My Work?

tav — Sat, 24 Jul 2010 02:55:35 -0000

Hey Sepp,

Thanks for the constructive criticism. I've put together a more plain language explanation:

* http://tav.espians.com/peer...

I'd really appreciate it if you could take a look at it and let me know if it's any clearer. Unfortunately there are so many many different aspects to the overall vision that it's hard to summarise it plainly and briefly.

Apologies in advance if my attempt isn't satisfactory — but I'd really love to know how I can make it better.

— Thank you, tav

Re: 4 Features To Make GitHub An Awesome Platform

tav — Thu, 08 Apr 2010 03:13:33 -0000

Hey Erik,

I agree that a truly RESTful API would be great — but there is the issue that not all browsers implement support for HTTP verbs beyond GET/POST... =(

This might be one of the reasons why GitHub's existing API only uses those verbs?