Martin Kleppmann's Blog - Latest Comments

Re: Building Go Test It: Fun with Scala and REST APIs

Dave — Thu, 19 Jan 2012 09:30:19 -0000

Cheers for the run down, going through it at the moment. Just to say that your POM above should use camelcase for modelVersion, groupId & artifactId — IDE is giving out =)

Best,
Dave

Re: Valuation caps on convertible notes, explained with graphs — Martin Kleppmann‘s blog

Davidjaymor — Mon, 16 Jan 2012 20:33:35 -0000

in case you are looking for a legal template of a convertible note with a cap, check out www.mystartuplegaltemplates.co...

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

Mail — Tue, 06 Dec 2011 17:18:13 -0000

An accounting book is exactly what I was NOT looking for. This article gives me a basic understanding I need to to start doing basic stuff. I'm not interested in "more advanced thinking" if that is required I'll do some more reading or go to an accountant.

This article is great and gives me a good starting point without having to read some long-winded accounting book where I would get completely lost.

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

Josh Haberman — Tue, 18 Oct 2011 00:16:23 -0000

I think you may be confused about the difference between understanding a problem deeply and knowing lots of incidental details.

The key insight, which the article captures brilliantly, is that double-entry accounting is equivalent to a directed graph. The specific procedures people have developed for checking accuracy (reconciliation) or for patching up mistakes in the graph have absolutely no bearing on the underlying model.

Your comment is like saying that the most important thing about Newtonian Mechanics is learning how to double-check your measurements (and how to correct bad measurements). Important in practice? Sure. But utterly irrelevant to the underlying theory.

Re: Load/performance testing a Rails application with ApacheBench

Alexander Plutov — Wed, 12 Oct 2011 05:07:03 -0000

Good article about ApacheBench http://plutov.by/post/apache_b...

Re: About/Contact — Martin Kleppmann‘s blog

Ashburngroup — Mon, 03 Oct 2011 14:42:22 -0000

Have you published a book about REST API in SCALA?

Re: Ruby on Rails vs. Java Enterprise

Gers32 — Sun, 18 Sep 2011 11:53:21 -0000

So it's three years later. Ruby on Rails has greatly improved, and Grails is now a contender. What framework did you end up choosing?

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

Martin Kleppmann — Fri, 22 Jul 2011 20:21:26 -0000

Jens, I'm sure the metaphor isn't for everybody, but it worked for me. If you're implementing accounting software, you'll probably find yourself representing the information pretty much in a graph form though, because it is equivalent to standard double-entry bookkeeping: the two entries are the two nodes that are connected by an edge (where the edge is a transaction and the nodes are accounts).

For example, you could implement an accounting system in a spreadsheet using one row per transaction, and the following table columns: date, description, amount, credit account, debit account. Then each row is an edge in your graph, and the credit and debit accounts names are the start and end of the edge, respectively.

Re: OAccounts - Setting your accounts data free

Martin Kleppmann — Fri, 22 Jul 2011 20:16:04 -0000

Hi Jens, unfortunately I've not had time to push it any further. I still think it's a good idea, but it probably needs to be taken on by someone closer to the accounting industry — I'm too far removed!

Re: OAccounts - Setting your accounts data free

Jens — Thu, 21 Jul 2011 18:38:24 -0000

What happened to this project?

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

Jens — Thu, 21 Jul 2011 16:51:00 -0000

A graph-theoretical approach would make sense if there were any algorithmical problems in accounting - but there aren't. Coloring nodes in a graph is the wrong metaphor to explain accounting to me. How will it help me to build an accounting software?

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

Luca Raviloli — Sun, 22 May 2011 00:20:52 -0000

Expanding your model to match how bookkeepers do it.

Make a profit and
loss node. Make an arrow from each of your blue nodes to the profit
and loss node transferring all the values form those nodes to the
profit and loss node. Now the blue nodes are zeroed.

Now profit and
loss node and it's subsidiary temporary accounts are subsidiary
accounts(nodes) to the pink node, Investor's capital. Do your income
statement from there. The edges to the profit and loss show the flows from accounts.

Then take the total in the profit and loss node
and make an edge for the whole amount to the Investors' capital
node(pink), or make a new pink node, Retained earnings account and pour the profit and loss into that. Then that will give the Balance sheet. So what is left
over or profit goes to the investors. Using one of your graph rules
split the capital nodes (pink) from the green node (blue are zero) and the
two sides are equal and opposite signed because if you add the two
groups you get zero thus making a proper balance.

In your model you may not want to clear profit and loss if you want to
make that a permanent account of capital. The book keepers put P/L in
and Investors' Capital account called Retained earnings after they
removed the dividends.

Blue nodes are temporary balances and are eventually transferred to Investor's Capital. That is why the sign is different. It represents money belonging to or owed to the Investors, from the accounting entitie's point of view.

No sign flipping? Investor's Capital is the mirror of
Assets-Liability. A-L=IC, (this is the clearest formula). Or Investor's capital and Liabilities are the compeating claims
on the assets; investors capital gets the remainder. Or, like your graph, A-L-C=0.

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

Southwest Schools — Fri, 20 May 2011 06:04:54 -0000

Nice explanation. Thanks for sharing.

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

Martin Kleppmann — Wed, 30 Mar 2011 03:25:29 -0000

In both a purchase and a sale, the extinguishing of the liability
happens by money or goods changing hands. In simple cases (buying a
bagel), the purchase and the payment happen at the same time, in which
case it never appears as an asset or a liability. However, if you
receive prepayment for a sale, that's a liability until you deliver
the goods/services. If you purchase something, but pay later, that's a
liability until you actually hand over the money.

Conversely, if you make a sale but allow the buyer to pay later, that
sale is an asset until you receive the money (at which point the money
becomes the asset). And if you pre-pay for a purchase, e.g. pay rent
in advance, that's an asset until you have actually received the
goods/services.

If you look in the graphs above at the way I split the sales into a
sale transaction and a payment transaction, that's exactly what's
going on here.

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

Esto — Tue, 29 Mar 2011 06:02:34 -0000

If a sale *is* a liability, what happens when the sale is completed (goods or sevices delivered) how do we record that this "liability" has been extinguished? And if we view sales as a liability, then logically expenses would be an asset - which just doesn't make sense.

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

Martin Kleppmann — Sun, 27 Mar 2011 17:51:05 -0000

Thanks! Sure, feel free to use it under Creative Commons http://creativecommons.org/lic...

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

accounting for programmers — Sat, 26 Mar 2011 14:12:19 -0000

Martin,

As discussed, we at www.accountingtutor.org have posted a piece to this article. Full credit has been given to you and a link provided. Great stuff. Can we use the entire article?

Re: Accounting for Computer Scientists — Martin Kleppmann‘s blog

wanorris — Tue, 22 Mar 2011 17:38:28 -0000

In a sense, a sale *is* a liability. A sale is the obligation to deliver goods and/or services in return for payment. A received payment is an asset, an undelivered good or service is a liability. If you receive prepayments from customers and have not delivered promised goods and services, the requirement to deliver such a good or service should appear as a liability to balance out the asset (the cash) that you received.

Or at least I think so. It's been 15 years since I took a year of accounting, so I could be mistaken. BTW, those courses have been quite useful at various times in my software development career. Even if you never start your own business, knowing how businesses work can be incredibly useful.

Re: What's so special about Y Combinator? — Martin Kleppmann‘s blog

Martin Kleppmann — Wed, 16 Mar 2011 02:54:04 -0000

It depends what you want. Most businesses are small and steadily profitable — think corner shops, restaurants, hairdressers etc., and that's a perfectly fine choice. However, that's not the kind of business that YC (or any venture investor) is interested in. The way how investors make money is from a fast-growing businesses which either generates a big exit, or becomes profitable at a large scale (more than $100m revenue per year). Some entrepreneurs find that an exciting prospect, others prefer organic, steady growth. You just need to figure out which one is your style.

Re: What's so special about Y Combinator? — Martin Kleppmann‘s blog

isomorphisms — Wed, 16 Mar 2011 01:39:50 -0000

> $ isn't a big draw unless you're 22 and have never made anything before

My thoughts exactly and this was a major turn-off about YC. (Suggests lack of seriousness of applicants.) But it could buffer the cost of moving to Silicon Valley.

Re: What's so special about Y Combinator? — Martin Kleppmann‘s blog

isomorphisms — Wed, 16 Mar 2011 01:37:33 -0000

Max, can you name a name?

Re: What's so special about Y Combinator? — Martin Kleppmann‘s blog

isomorphisms — Wed, 16 Mar 2011 01:34:55 -0000

Although, Martin, is it really true that the only desirable outcome is a big exit? One could also seek to craft a business that will bring in steady income without too much management work (or at least brings in enough money that you can hire people to solve the ongoing problems).

Re: What's so special about Y Combinator? — Martin Kleppmann‘s blog

Ryan Hoover — Tue, 15 Mar 2011 17:59:31 -0000

Great write-up, Martin. I've found YC alumni extremely supportive and willing to help YC applicants, not just their fellow peers.

Re: What's so special about Y Combinator? — Martin Kleppmann‘s blog

isomorphisms — Tue, 15 Mar 2011 14:37:53 -0000

Question that needed to be answered. Thanks.

Re: What's so special about Y Combinator? — Martin Kleppmann‘s blog

Martin Kleppmann — Tue, 15 Mar 2011 13:19:33 -0000

I've not heard any stories of "PG becoming disinterested". Sure, not
every startup stays close to YC as it journeys on; some call back
frequently, some occasionally, some never. But as far as I can tell
that's entirely the choice of the founders. YC is there if you want
them, since they want everyone to succeed, but they never force
founders to do anything.