Tom Moertel's Weblog: Category good stuff

A coder's guide to usage guides (and why you ought to use one)

Tom Moertel — Wed, 19 Jul 2006 12:55:00 -0400

Note for coders: If you’re one of the fine folks who reads my blog for the coding content, I’ll be up-front with you: this article represents about 2,500 words of non-coding text. Nevertheless, you ought to read it because I am writing about something that’s important to you.

We coders, if you stop to think about it, spend our lives doing some insanely complex and seemingly bizarre stuff. If we want the rest of humanity to understand us and what we do, we must be prepared to explain our world to them; they are not going to figure us out on their own.

Thus the burden of making our world understandable and relevant to others rests upon our own shoulders. To carry the burden, we must be able to communicate complex ideas, share the full depth and beauty of our creations, and help others see that the things we devote our lives to are worthy and fascinating. In sum, we need to invest in our writing skills. Writing is still the most effective, scalable means we humans have for capturing and sharing complex knowledge. If we want to reach the rest of the world, we must become good writers.

Unfortunately, writing well is difficult. Like programming, writing is a struggle. If you want to improve, you must struggle often. (That’s why I blog.) And while I cannot struggle for you, I can tell you about an unappreciated writing tool that has become my go-to reference when writing. It has made my struggles a bit less difficult and a bit more fun. That reference is not the ever-popular dictionary.¹ Rather, it is the usage guide.

Usage guides are designed to explain the finer points of using our language and, if they’re good, they offer sensible advice. Have you ever wondered whether to use affect or effect, compliment or complement, farther or further, compose or comprise, less or fewer, precede or proceed? Do you know why you shouldn’t confuse enormity for enormousness, masterful for masterly, or purposely for purposefully? How should you approach who and whom? And how should you address the thorny issue of sexist language? These are the kinds of questions that usage guides were created to answer.

A good usage guide isn’t afraid to prescribe advice. Some more-recent guides, however, are less prescriptive and more descriptive, following the lead of modern dictionaries. Instead of offering advice on tricky writing issues, these guides describe the many viewpoints in play and let you choose among them. I think this weakens a guide, and so my recommendations below are biased toward prescriptive guides.

Unlike dictionaries, usage guides are not plentifully stocked in book stores. On a recent visit to my local Barnes & Noble, for example, I was surprised to learn that the reference section contained no usage guides at all. For this reason, I will provide Amazon links to the guides I like.

Let us begin our tour of usage guides at the beginning, with Fowler.

Fowler’s Dictionary of Modern English Usage

Of the authors of usage guides, no one is more revered than Henry Watson Fowler. He and his brother Francis wrote The King’s English, a reference published in 1906 that provided lengthy advice on issues of style and grammar. Afterward, the brothers Fowler collaborated on their next major work, but in 1918, after eight years of planning, Francis died, leaving Henry alone with the daunting task of writing the book, which ultimately required another eight years. Published in 1926, this book was to become the influential masterwork, A Dictionary of Modern English Usage, most often called Fowler’s Modern English Usage or simply Fowler.

Since its publication, Fowler has become the quintessential prescriptive guide to English usage and is still beloved among picky writers. One can’t help but respect the author, who doesn’t cower behind dry, academic language but instead throws himself vividly onto every page and seems to go out of his way to fight pedantry.

The followers of Fowler prize his precision but love him for his quirky voice. When you read Fowler, you get Fowler. Consider the first edition’s entry on superiority (remember, this was written almost a century ago):

Superiority. Surprise a person of the class that is supposed to keep servants cleaning his own boots, & either he will go on with the job while he talks to you, as if it were the most natural thing in the world, or else he will explain that the bootboy or scullery-maid is ill & give you to understand that he is, despite appearances, superior to boot-cleaning. If he takes the second course, you conclude that he is not superior to it; if the first, that perhaps he is. So it is with the various apologies (to use an expressive colloquialism – if we may adopt the current slang – . . . ) to which recourse is had by writers who wish to safeguard their dignity & yet be vivacious, to combine comfort with elegance, to touch pitch & not be defiled. They should make up their minds whether their reputation or their style is such as to allow of their dismounting from the high horse now & again without compromising themselves; if they can do that at all, they can dispense with apologies; if the apology is needed, the thing apologized for would be better anyway. . . . (Fowler, 1st ed., 1926.)

The editors of Fowler’s later editions, to the disappointment of Fowlerians everywhere, have muted Fowler’s voice while undertaking the necessary work of making the reference more accessible to contemporary readers and writers. For example, the superiority entry from the 2nd edition, edited by Sir Ernest Gowers, reads as follows:

Superiority. Much misplaced ingenuity in finding forms of apology is shown by writers with a sense of their own superiority who wish to safeguard their dignity and yet be vivacious, to combine comfort with elegance, to touch pitch and not be defiled. Among them are: To use and expressive colloquialism – in the vernacular phrase – . . . . (Fowler, 2nd ed., 1965)

Most writers consider Gowers’s stewardship of Fowler completely respectable. The second edition still speaks with Fowler’s voice, and Gowers’s revisions make the work more practical for modern writers. The second edition is seen in much the same light as the screenplay adaption of Tolkien’s Lord of the Rings: the original was changed, but the changes were understandably necessary and, most important, the integrity of the original was largely preserved.

The third edition of Fowler, however, is the subject of controversy. Edited by Robert W. Burchfield and published in 1996, The New Fowler’s Modern English Usage differs from the previous editions in two important respects. First, it no longer speaks with a distinctly Fowlerian voice. Second, Burchfield allows descriptivism to creep into the celebratedly prescriptive reference.

Nevertheless, it is a useful reference work. Even the third edition’s detractors admit that it is a perfectly respectable, modern usage guide. Their main complaint is that the work is no longer Fowler, and I am inclined to agree. Burchfield distances himself from Fowler, as the third edition’s entry on superiority demonstrates:

Superiority. Fowler’s term for the use of a slang expression or a socially divisive remark preceded by a distancing or defensive comment implying that in normal circumstances the speaker would not deign to use such an expression himself or herself. Such distancing remarks include as they say; if the word may be permitted; . . . . (Fowler, 3rd ed., 1996)

Burchfield writes that superiority is “Fowler’s term,” putting Fowler in the third person instead of adopting Fowler’s voice. Instead of recommending a position – taking a stand – he tells you what positions others have taken (in this case Fowler himself). He observes and describes. There’s nothing inherently wrong with this approach, but in a usage guide, I prefer to know what the author himself thinks.

I suspect that many of the third edition’s critics would have welcomed the book had the Oxford University Press not used Fowler’s name in its title. But the OUP did call it Fowler, and the reviewers judged it accordingly. Nevertheless, the third edition is a worthy reference, and I use it second only to Garner, which I will get to in a moment.

Because all three editions of Fowler have their proponents, all three editions are still available today. The first edition was out of print for some time (I found my copy, the original 1926 letterpress edition, at a book sale), but the third edition’s awakening of the Fowlerian hordes led to renewed interest in republishing the earlier versions. Now all three are readily available:

Which version is for you? If you want an up-to-date, practical reference, get the 3rd edition (or, better yet, get Garner, discussed next). If you want less descriptivism and more of the original Fowler flavor, get the 2nd edition instead. You probably should not get the first edition unless you are fascinated by the English language or want to experience the unadulterated, original Fowler, in which case you are the kind of person who probably has all three editions already.

Garner’s Modern American Usage

My favorite book on usage is Garner’s Modern American Usage. Bryan Garner, a pragmatic prescriptivist, has crafted an excellent usage guide of a distinctly Fowlerian flavor. (That he also founded the H.W. Fowler Society ought to tell you something about the way he approaches usage.)

In Garner, the detractors of the third edition of Fowler have found a successor to Fowler in spirit. Where Burchfield describes, Garner prescribes. Consider how the authors approach the subject of sexist language. Burchfield writes:

sexist language. 1. As indicated in numerous articles in this book . . ., feminists and others sympathetic to their views, from about the 1970s onwards, have attacked what they take to be male-favouring terminology of every kind and have scoured the language for suitable evidence and for gender-free substitutes. Their argument hinges on the belief that many traditional uses of the language discriminate against women or render them ‘invisible’ and for these reasons are unacceptable. The various types of alleged linguistic discrimination need not be repeated here. Perhaps the most obvious reference works on the subject are . . . . 2. Some landmarks. As rough indicators of the development of feministic views on gender-free language it might be useful to set down in chronological order details of some decisions and discussions that have taken place in various English-speaking countries since 1988: . . . . (Fowler, 3rd ed., 1996)

Burchfield then devotes the next two subsections (3 and 4), representing the balance of the full-page entry, to examples from both sides of the debate and to describing the academic viewpoint on the subject.

In contrast, Garner distills the problem to its essence and offers practical advice:

SEXISM. A. Generally. If you start with the pragmatic premise that you want to avoid misleading or distracting your readers, then you’ll almost certainly conclude that it’s best to avoid sexist language. Regardless of your political persuasion, that conclusion seems inevitable – if you’re a pragmatist.

But does avoiding sexism mean resorting to awkward devices such as he/she? Surely not, because that too would distract many readers. What you should strive for instead – if you want readers to focus on your ideas and not on the political subtext – is a style that doesn’t even hint at the issue. So unless you’re involved in a debate about sexism, you’ll probably want a style, on the one hand, that no reasonable person could call sexist, and on the other hand, that never suggests you’re contorting your language to be nonsexist.

(Garner, 2nd ed., 2003)

Garner then goes on to provide two pages of practical suggestions:

B. The Pronoun Problem. . . .
C. Words with man- and -man. . . .
D. Differentiated Feminine Forms. . . .
E. Equivalences. . . .
F. Statute of Limitations. . . .

(Garner, 2nd ed., 2003)

Garner also provides a bibliography (subsection G) should you wish “to inquire further into this interesting subject.”

Most of the differences between Garner and Fowler 3 are not so pronounced. Either reference would serve the needs of most writers. Nevertheless, I find myself reaching for Garner first and then, if I want a second opinion, for Fowler 3. So my top choice is Garner.

This ends my brief tour of usage guides. I have overlooked many useful guides, such as Merriam Webster’s, but I don’t find them as useful as Garner or Fowler. If you happen to be in a good library, prowl the reference section and draw your own conclusions.

My advice: if you don’t have a usage guide, get one and use it

If you care about writing enough to own and use a dictionary, you probably ought to have a good usage guide, too. While there are many usage guides available, if you can have only one, my recommendation would be for Garner’s Modern American Usage. If you are the kind of person who likes the luxury of a second opinion, complement Garner with The New Fowler’s Modern English Usage, 3rd ed. (Paperback) (or 3rd revised ed. in Hardcover). Finally, if you want to experience Fowler unaltered, pick up a copy of the first edition of Fowler. If you find it at a book sale, you might be able to score a copy of the original 1926 printing.

Even if you don’t want to buy a usage guide, you ought to sign up for Garner’s Usage Tip of the Day. Every day you’ll receive an emailed tip from Garner and maybe a quotation or two on writing. Sign up at the OUP’s email-subscription page.

I prefer printed reference books, but if you would rather get your usage information online, here are a couple of starting points:

There you have it: the usage guide – my favorite writing tool. If I haven’t convinced you to add one to your own writing toolkit, at least flip through Garner or Fowler the next time you’re in the library. Maybe then you’ll reconsider. For me, usage guides are a no-brainer: anything that makes writing a bit more fun and a bit less of a struggle is worth having on my bookshelf.

Because dictionary publishers have made the dictionary-and-thesaurus duo a part of every college freshman’s standard gear, I will assume that you already have a good dictionary and so focus my attention exclusively on usage guides. If I’m wrong about this, I can recommend my favorite dictionary, The New Oxford American Dictionary. (Please do not bother to inform me that the NOAD is inferior to The Shorter OED because if that’s the way you roll, you clearly don’t need my advice.) Another good choice is Merriam-Webster’s Collegiate Dictionary, 11th Edition.

Good stuff: Tab Mix Plus for Firefox

Tom Moertel — Fri, 03 Mar 2006 10:13:00 -0500

It’s a real forehead-slapper that I only now started using Tab Mix Plus, a Firefox Extension that lets you completely customize Firefox’s tab behavior. It has about a billion settings, but the defaults are carefully chosen so that it does good stuff without needing adjustment. Unless you’re a goofball, you will notice an immediate improvement in browsing.

Among the more useful features are:

Session management – If you quit Firefox (or it crashes), Tab Mix Plus can restore your session when you restart.

Sensible forking – If you “fork” a tab, the new tab appears next to its parent, not at the end of the tab bar. As a result, related tabs are clustered, not scattered, and the ordering of tabs corresponds with your brain’s picture of the session. (Also, you can drag tabs, should you want to impose another ordering.)

I’m-new highlights – Tabs you fork into the background are highlighted to let you know they contain new, unseen content. This is great for power sessions when you otherwise couldn’t remember which tabs you’ve already seen.

Locking/Protecting – Lock a tab, and the browser won’t change it. If you click a link or do anything else that would replace the contents of the tab, it will fork a new tab instead. Protect a tab, and you can’t accidentally close it; you must unprotect it first.

If you use Firefox, you need to check out Tab Mix Plus. It’s good stuff.

Good stuff: Aldo Coffee Company

Tom Moertel — Mon, 30 Jan 2006 21:57:00 -0500

I love espresso. It’s my favorite way to enjoy coffee. Even so, I almost never order espresso in coffee shops because, here in the United States, very few coffee shops have mastered the exacting process by which espresso is made. Dr. Josuma John of the Josuma Coffee Company writes that “more than 95 percent of North American espresso is poorly made, and, in fact, undrinkable.” My experience with Pittsburgh-area coffee shops in the last decade provides no evidence to refute Dr. John’s claim.

If espresso in the United States is so bad, why do Americans drink enough of it to support a Starbucks on every street corner? The reason is that Americans drink espresso almost exclusively in the form of milk-based beverages: cappuccinos, lattes, and mochas. Milk and flavored syrups are the main attractions. Espresso serves only as a coffee-flavored backdrop in which bitterness, a characteristic of poorly made espresso, complements the abundant sweetness of milk laced with sugar syrups. American coffee-shop owners thus have little incentive to offer better espresso to their customers – bad espresso is good enough.

Because of this sad reality, I have developed through hard experience the following reliable guideline for ordering espresso at American coffee shops: Don’t. The one exception I make is for new coffee shops, at which I will try a double espresso, just to see what I get. Almost always, I get a bad espresso, bitter and watery.

And that is what I had expected back in April 2005, when I spotted the brand-new sign for Aldo Coffee Co. in my home town of Mt. Lebanon, Pennsylvania, located in Pittsburgh’s South Hills. I went in, dragging my wife along, and placed my order.

Then something unusual happened. The barista asked me, somewhat expectantly it seemed, if I drank espresso regularly. When I responded, “Oh my, yes,” she seemed pleased. When she followed up by asking me if I read alt.coffee, I was stunned. When I observed that she was timing my shot, my brain actually shut down for a few seconds while it forcibly recalibrated itself to accommodate the seemingly impossible: that I was standing in a coffee shop in my home town, conversing with a barista about alt.coffee, and mere seconds away from receiving what was very likely to be good espresso.

Good espresso

And, in fact, the espresso was good. Aldo uses Intelligentsia Coffee & Tea’s Black Cat Espresso Blend, which is a great espresso blend, skillfully roasted. It seems to work well in commercial espresso environments, where extraction temperatures can be carefully controlled. (I think that Black Cat tips into burnt flavors when pulled too hot.) All in all, a delightful cup.

Since then, I have visited Aldo Coffee Co. regularly, typically at least once per week, and the espresso has been reliably good. I would say that out of five cups, I typically get one that is pretty good, two that are definitely good, one that approaches greatness, and one that is great. I don’t recall ever having received an undrinkable cup. At one time (when their machine was running hot or Intelligentsia was off their target profile) I did receive a stretch of cups with emphasized burnt flavors, but still the overall impression was pretty good.

Certainly, there is room for improvement, but with espresso, that is the way it goes. Espresso is a tricky beverage, requiring continual fine tuning and practice. To put Aldo’s performance into perspective, at home, where I can focus on a single customer’s tastes – mine – and I am working with equipment that I have practiced on for about a decade, I don’t do much better: out of five tries I get one pretty good, one good, two approaching greatness, and one great. (One of the great shots is pictured in my blog’s banner.) At most coffee shops, five out of five times I get something not worth drinking.

The only other contender in Pittsburgh is La Prima Espresso, which gets the nod of approval from Tea Leaves guys, whose tastes I tend to agree with. My experience at La Prima, however, has not matched theirs. More often than not, the espresso I have been served was over-extracted. I have received a good cup on occasion, maybe one for every five visits, but for the most part I have left disappointed.

All of this leads me to the following claim: Aldo Coffee Co. is the most reliable place I know of to get good espresso in Pittsburgh.

Good service and food

That Aldo’s owners are willing to invest in making good espresso, a tricky product that I suspect less than five percent of their customers appreciate, suggests they are atypically attentive to all of their offerings. After all, if you are going to cut corners, why not start with espresso?

The Aldo staff seems to bear out this theory. Everybody is friendly, and several baristas know me by name, despite that I’m not a daily customer.

Their sandwiches also point to something beyond the norm. I have tried the “tuna & artichoke” and the “eggplant, prosciutto, & mozzarella,” and both were delicious, reminding me of the typical fare in Italian bars, where they care about these things. The breads (from Mediterra Bakehouse) are crusty and flavorful, balancing with the fillings, neither dominating the other.

I am not a dessert person and avoid coffee-shop sweets, but I do enjoy Aldo’s sfogliatelle. Too much. ’Nuff said.

Good web site

One more indicator that Aldo is not your typical Pittsburgh coffee shop is that they blog. And, unlike me, they keep their blog up to date. Their blog is a part of their web site at www.aldocoffee.com, which has menus, Pittsburgh-related tidbits, and even barista profiles, so you can get to know their staff. This, too, suggests an unusual attentiveness to things that other shops overlook.

It all adds up to good stuff

Good espresso, good service, good sandwiches, good sfogliatelle, and an informative web site featuring an oft-updated blog. Clearly, something unusual is happening at Aldo Coffee Co. I am not sure how they make it happen, but I am sure about one thing: Aldo is good stuff.

Update 2006-03-06: I originally wrote that Black Cat was “darker than a traditional Northern Italian roast.” I just got some more Black Cat, and this batch is lighter than I had remembered. It looks to have just brushed up against the second crack.

Good stuff: Google Local for mobile devices

Tom Moertel — Mon, 28 Nov 2005 18:14:00 -0500

While traveling in unfamiliar territory over the Thanksgiving holiday, I needed coffee. Bad. So I fired up my mobile phone’s web browser and surfed to Google Mobile. Just as I was about to submit a search to the Google Local web form, I noticed a link to something intriguing: Google Local for mobile devices (GLM).

Curious, I checked it out. I discovered that GLM is a Java applet that runs on your cell phone to provide a phone-optimized interface to Google Local’s servers. After a quick over-the-air download, I was ready to try it out.

I fired up the GLM applet and entered my query: “coffee elkin,nc.” In a few seconds, I had a clean, crisp map that showed a number of coffee shops. Pressing 1, 2, 3, and so on, I got the details for the various shops. When I found one I liked, I selected “Directions to here” from the details menu. Then I used my phone’s cursor keys to select a spot on the map that approximated my current location. After a few seconds of “Routing…” the map updated and plotted a path to coffee salvation. In a few minutes, I had a coffee in my hand and was back on the road.

The process was fast and painless. It just plain worked. I knew then and there that Google Local for mobile devices is good stuff.

Right now GLM is only available on Cingular and Sprint networks, and even then only for more-recent Java-based phones. [Update: In a comment, Mark points out that GLM works on some T-Mobile phones, too.] If my phone or network didn’t support GLM, I would seriously consider switching to a phone-network combo that did. GLM is that cool.

In any case, when your cell-phone contract is up, you ought to factor GLM into your renewal/switch plans. A cell phone with GLM is a heck of a lot more useful than one without it.

Good stuff: Foyle's War

Tom Moertel — Thu, 01 Sep 2005 15:20:00 -0400

“Reality” shows have plunged mainstream television into an entirely new depth of stupidity – and for television, that’s saying something. Fortunately for us, some programs defy the downward trend, and Masterpiece Theatre’s Foyle’s War is one of the best.

Foyle’s War follows Christopher Foyle during World War II in England. Germany occupies mainland Europe, and invasion seems imminent. Ordinary people carry on as best they can, but an undercurrent of desperation pulls at them, and some give in to terrible urges, including murder.

Foyle, the Detective Chief Superintendent of the Hastings police department, sees the worst of it. Denied by his superiors the right to join the armed forces, Foyle fights a different war, one that claws away at his home town. Hastings, on the south coast of England, is a likely invasion site, and its citizens must live with barbed wire, road blocks, anti-invasion exercises, and increasingly dire rumors. As crime becomes more frequent and more brutal, Foyle and his his small staff – Detective Sergeant Paul Milner and his driver Samantha “Sam” Stewart – struggle against what threatens to become anarchy.

As you may have guessed, Foyle’s War is of the mystery genre. I am not one of those “mystery persons,” but Foyle’s is so far beyond what is typical of the genre that I classify it simply as Good Stuff. This is not Murder, She Wrote. The writing is great. The acting is subtle – Michael Kitchen conveys more meaning with a single nod than most actors do with their dialog. The sets, camera work, and directing collaborate to draw the helpless viewer into WWII-era England. The result is fun television that need not be considered a guilty pleasure.

So far, four series of feature-length Foyle’s War episodes have been filmed. The first two are out on DVD now in the USA (see links below). If you have Netflix, you know what to do. The third will be released on DVD in November, but its episodes will be airing Sundays on PBS’s “Mystery!”, 11 September – 2 October. (Check your local listings to make sure you don’t miss it.) The fourth series is still in production.

If you are sick of reality television, you now know the cure.

Amazon Links

[If you buy something via one of those links, Amazon will give me a small referral fee, approximately one zillionth of what I could have earned had I been working instead of playing around on my blog. Let that be a lesson to me.]

Good stuff: One-Touch high capacity stapler from Staples

Tom Moertel — Tue, 31 May 2005 12:00:00 -0400

I read a lot of academic papers, most of which are distributed in Postscript or PDF format. Because they are usually typeset for publication in journals, reading them on the computer is fatiguing. The type is too small, and I must zoom and scroll to compensate.

For this reason, I almost always dispatch the papers to the laser printer, staple the resulting stack of pages, and then read the printed version. Not only is this easier on the eyes, but also I can read while reclining in a comfortable chair, preferably with an espresso on the side table. I can also scribble notes in the margins. It’s a good deal.

But there’s a catch: the system only works for short papers. If more than fifteen pages come out of the laser printer, the weak link is readily exposed: the everyday desk stapler. Try to bind a non-trivial stack of pages, and you’ll soon have bent staples, mangled pages, and a stapler that is flat-out jammed.

Nevertheless, I can go further than most. For I have learned the dark arts by which desk staplers can be made to bind beyond their mortal limits. Through cunning manipulation and fell incantation (swearing, mostly) I can coerce an ordinary desk stapler into binding perhaps twenty pages. But that is the utmost limit; beyond, there is only suffering.

That is, until today. For today I walked into my local Staples and saw the One-Touch high capacity stapler. Half stapler, half hammer tacker, this bad boy had my name written all over it. The price tag was a suspiciously high $29.98, but I didn’t care.

Sold.

How does it work?

The central idea is that by compressing the “one touch” lever, you store energy in a spring, building up a sizable potential. At the end of the stroke, this potential is released as a sudden impulse. The impulse in turn breaches the fabric of space and time, opening an extra-dimensional gateway through which demon elves enter our plane of existence and lovingly drive the foremost staple through the stack. In an instant, the deed is done, the gateway closes, and you are left with a perfectly stapled stack of pages (and the lingering scent of brimstone).

That’s the theory. But does it work in practice?

Putting it to the test

Test number one: A 27-page paper on survival analysis. I aligned the pages and placed the corner (at a perfect 45-degree angle, naturally) into the maw of the stapler. I depressed the lever. Slowly, energy built until – Thwack!

A perfect staple.

But that was a paper on statistics. Easy. How would the stapler fare on computer science papers, which are notoriously difficult to staple?

Test number two: “An Embedded Domain-Specific Language for Type-Safe Server-Side Web Scripting.” This paper was especially tricky. First, it was printed on 24# premium-grade, hard-calendared stock. (When I read comp-sci papers, I read in style.) Second, a long time ago I had tried to staple this very same paper with a typical desk stapler, which had failed. In order to get the original staple to take, in fact, I had to bend its points into the paper by hand, using a penny. How humiliating!

Now I tested the “One Touch” on this formidable comp-sci paper. First, I removed the old, pathetic staple. Then I re-aligned the pages and placed them into the One Touch. I pressed down on the lever and – Thwack!

Another perfect staple.

Conclusion: This stapler is the real deal.

If you read mighty documents, you deserve a mighty stapler. And the One Touch is a mighty stapler. Keep that in mind the next time you are roaming the aisles of your local Staples store.

Cool stuff: Composable memory transactions

Tom Moertel — Sat, 09 Apr 2005 14:10:00 -0400

If you write software that deals with concurrency, you are doubtless familiar with the painful limitations of the concurrency abstractions that most programming languages, runtimes, and operating systems offer us humble programmers. The one-big-select idiom, the need to impose a global ordering policy on lock taking, and the myriad things that can unexpectedly bite you in the behind when managing threads are not-so-subtle reminders that the programming world still has a few fundamental problems to solve.

Thus I was impressed when I read Tim Harris, Simon Marlow, Simon Peyton Jones, and Maurice Herlihy’s paper on Composable memory transactions a couple of months ago. The paper introduced some Very Cool Stuff (especially if you program in Haskell, for which there is an implementation available). More recently at the Links meeting at ETAPS (another cool thing a’brewing), the same team gave a talk on the subject: Concurrency Unlocked: transactional memory for composable concurrency. Check out the slides from the talk for a summary of the problem and the STM solution.

The gist is that today’s ubiquitous concurrency abstraction – the lock – is fundamentally at odds with the most successful technique we humans have for building complex systems: gluing simple systems together. Composable memory transactions, on the other hand, do not have this problem. As a result, they offer a fundamentally simpler and more mentally scalable solution for building complex concurrent systems.

To quantify this coolness, we have only to look at section 4.2 of the paper. In about 25 lines of code the authors give an implementation for what is effectively the heart of an instant messaging server. I am not kidding. Multiple writers with serialized writes into each channel, multiple readers on each channel with independent read positions and buffering. Yeah, it’s in there.

Do yourself a favor. Check out the slides from the talk and then read the paper. This is some seriously cool stuff. You ought to know about it.

Update 2007-01-16: Since I wrote this article, Software Transactional Memory has received a lot of attenttion. Here are a couple of pointers worth checking out:

Beautiful Concurrency (PostScript paper). An approachable introduction to STM that doesn’t assume you already understand Haskell.
Wikipedia’s entry on Software Transactional Memory is pretty good, too.