@hackage penny-lib0.4.0.0

Penny is a double-entry accounting system. It is inspired by, but incompatible with, John Wiegley's Ledger, which is available at http://ledger-cli.org/. You will want to install the package penny-bin by using cabal install penny-bin, which will install the executable program and this library as well. Features:

• Penny is a double-entry accounting system. It uses traditional accounting terminology, such as the terms "Debit" and "Credit". If you need a refresher on the basics of double-entry accounting, pick up a used accounting textbook from your favorite bookseller (they can be had cheaply, for less than ten U.S. dollars including shipping) or check out http://www.principlesofaccounting.com/, a great free online text.

• Penny is based around Penny.Lincoln, a core library to represent transactions and postings and their components, such as their amounts and whether they are debits and credits. You can use Lincoln all by itself even if you don't use the other components of Penny, which you may find handy if you are a Haskell programmer. I wrote Penny because I wanted a precise library to represent my accounting data so I could analyze it programatically and verify its consistency. I wrote it in Haskell not because I wanted to write something in Haskell but because Haskell is the best tool for this job (I used to use the shell, combined with Ledger, which is a messy combination.)

• Penny's command line interface, Zinc, and its reports, Cabin, give you unparalleled flexibility to filter and sort postings. Each posting within a transaction may have its own flags assigned (e.g. to indicate whether the posting is cleared) and each posting may have infinite "tags" assigned to it, giving you another way to categorize your postings. For instance, you might have vacation related postings in several different accounts, but you can give them all a "vacation" tag.

• Full Unicode support. Also, you may set which characters you wish to use to represent the radix point and the digit grouping character in your ledger file.

• Penny's reports, in Penny.Cabin, have color baked in from the beginning. You do not have to use color, though, which is handy if you are sending output to a file or if, well, you just don't like color.

• Penny's reports are customizable in Haskell, giving you an easy and powerful way to make your own reports without writing cryptic formatting strings.

• Penny handles multiple commodities (for example, multiple currencies, stocks and bonds, tracking other assets, etc.) in an easy and transparent way that is consistent with double-entry accounting principles. It embraces the philosophy outlined in this tutorial on multiple commodity accounting: http://www.mscs.dal.ca/~selinger/accounting/tutorial.html.

• Penny stores amounts using only integers, building from the Data.Decimal library available at http://hackage.haskell.org/package/Decimal. This ensures the accuracy of your data, as using floating point values to represent money is a bad idea. Here is one explanation: http://stackoverflow.com/questions/3730019/why-not-use-double-or-float-to-represent-currency. The use of integer arithmetic also makes Penny simpler internally, as there is no need for arbitrary rounding to compensate for the bizarre and inaccurate results that sometimes arise from the use of floating-point values to represent currencies.

• Freely licensed under the MIT license. If you take this code, improve it, lock it up and make it proprietary, and sell it, AWESOME! I haven't lost anything because I still have my code and, what's more, then maybe I can buy your product and not have to maintain this one any more!

• Uses no GHC extensions. However, the code is only tested under GHC and for all practical purposes it will only run under GHC at this time because it uses libraries such as Data.Text that are available only under GHC. Despite this I expect I will continue to avoid language extensions.

Non-features / disadvantages:

• Written in Haskell. Yes, I think Haskell is the best tool ever, but its compiler is not as commonly installed as compilers for C or C++, and non-Haskellers will probably find Penny to be more difficult to install than Ledger, as the latter is written in C++.

• Handling commodities requires that you set up multiple accounts; some might find this cumbersome.

• Young and not well tested yet. Also, only tested on Unix. It probably would not be difficult to make Penny run on Windows; if someone wants to do that, go ahead.

• Full Penny functionality is available without a Haskell compiler; you could even use a pre-compiled binary. However, to fully customize Penny, you will need a Haskell compiler installed.

• Can be slow and memory hungry with large data sets. I have a ledger file with about 28,000 lines. On my least capable machine (which has an Intel Core 2 Duo at 1.6 GHz) this takes about 1.4 seconds to parse. Not horrible but not instantaneous either. Generating a report about all these transactions can take about seven seconds and a little less than 300 MB of memory. I have eliminated all the obvious slowness from the code and attempted a rewrite of the parser, which made no difference; other ideas to speed up Penny with large data sets would involve substantial changes and this is not at the top of my list because the program is currently usable with relatively recent hardware.

More details about the organization of the Penny modules is available by examining the top Penny module.

This is only a library. For the executable package, which also includes more documentation, search for the penny-bin package on Hackage.