A $10 Million Comma (Not Really)

April 2, 2017 —

There’s been a whole bunch of reporting (from The New York Times, for example) about a case in which a (deliberately?) omitted comma between two list items in a Maine statute might cost a dairy company $10 million. Even the US Court of Appeals for the First Circuit gives the humble comma star treatment here (first sentence of its opinion: “For want of a comma, we have this case.”).

This is not a case about a comma. No case is about a comma. The problems with Maine’s statutory drafting conventions (and probably statutory drafting conventions in general) go way beyond commas. If someone comes to you with a $10 million question that hinges on a comma, you’ve got a much bigger problem.

The powers that be in the State of Maine offer the Maine Legislative Drafting Manual. Its recommendations on lists don’t make much sense.

Complicated text. When several of the items in a series are more than one typed line long or a section has complex internal punctuation, the text should be written in outline form for ease of reading. In addition, outlining is often useful when a legislative sentence contains several cases, conditions or exceptions. Occasionally, however, readability may suffer as a result of outlining if the method is used too extensively. (p. 72, emphasis mine)

This echoes Reed Dickerson’s Fundamentals of Legal Drafting (which is cited a few times in the Maine Legislative Drafting Manual):

The draftsman should distinguish the outline form as an aid to analysis from the outline form as a means of expressing final text. The former must, in the interest of analysis, be severely hierarchical; the latter, in the interest of clarity, predominantly linear. Thus, an instrument whose apt concepts and logical sequence reflect a sophisticated, detailed working outline may finally appear merely as a sea of numerically coordinate sections. Even in the final text of longer and more complicated instruments, the hierarchical form must be used with restraint. (§5.1)

When we’re talking about statutes, how can readability suffer from using an outline list? Why use outlining with restraint? Statutes aren’t novels – no one is reading this stuff cover to cover. When would an outline list be anything but helpful?

Outlining sections. …begin each item with an upper-case letter, use a conjunction after the next to the last item, end each item except the last with a semicolon and end the last item with a period. (p. 73)

Why is all of this necessary? Is a line break really such a subtle way to show the end of a list item that you need semicolons and conjunctions?

Apple has created technical documentation for dozens, maybe hundreds, of devices and applications over the years. Here are Apple’s guidelines for punctuation in lists:

List items that are fragments or that complete the thought started by the main clause should not end with a period; list items that are complete sentences should end with a period.

That’s it. No terminal punctuation (unless the list items are complete sentences). No conjunctions ever. Nice and simple.

(“But how can anyone possibly know whether items in a list are conjunctive or disjunctive without an and or an or after the penultimate item?” you’re no doubt asking. Which of these options seems like the clearest, most user-friendly way to set up a conjunctive or disjunctive list: one word after the second to last list item, so that the reader has to puzzle through almost the entire list before understanding how it works; or a few words before the list?)

Indented and blocked paragraphs. A paragraph may apply to the entire section or to a previous subunit (subsection, paragraph, etc.), depending on its format. In order to eliminate any ambiguity in determining this relationship, the drafters of the 1964 Revision of the Statutes established the following rule to govern “indented” and “blocked” paragraphs. If the paragraph is unnumbered, unlettered and indented only on the first line, the paragraph applies to the entire section. If the paragraph is “blocked” (the first line is not indented, but the paragraph may have an indented left margin), the paragraph applies to the nearest previous subunit that shares the same margin or is located the same number of spaces to the right of the margin. (pp. 75–76)

Saying that what a paragraph applies to – what a paragraph means – depends “on its format” goes against decades of recommendations on technical writing (and statutes are pretty close to technical writing). In technical writing, you keep presentation (formatting) and content separate. This separation is one of the essential aspects of CSS, XML, and content management systems.

I fully accept that when you’re working with outline lists (as you often do in statutes), indentation (a formatting attribute) is essential to show hierarchical level (content). But a guideline that says that content depends on formatting gets the relationship backward: formatting depends on content.

This recommendation is also inconsistent with another one:

Flushing left. When you outline, do not use numbered subsections or lettered paragraphs in the middle of running text, a device commonly known as “flushing left.” Flushing left reduces readability and usually increases ambiguity.


1. Minor; Class E crime. A person who has not attained 20 years of age who:

A. Consumes alcohol in a public place; or

B. Operates a school bus without a valid driver’s license

commits a Class E crime.

The section is less confusing when written:


1. Minor; Class E crime. A person who has not attained 20 years of age commits a Class E crime if that person:

A. Consumes alcohol in a public place; or

B. Operates a school bus without a valid driver’s license. (p. 75)

I personally don’t see much difference in terms of readability and ambiguity between these two examples (I slightly prefer the first one because it’s closer to everyday English). I also can’t imagine an example in which there would be much difference.

But in any event, it’s strange that it’s OK to have stuff after a list when you call it an “indented” or “blocked” paragraph, but not when you call it “flushing left”.

Parenthetically, I don’t think it matters that “flushing left” happens mid-sentence. Whatever the sentence structure, what your brain internalizes is something like:

Blah blah blah:
• Item 1
• Item 2
• Item 3
Result or exception

On to commas:

Commas are probably the most misused and misunderstood punctuation marks in legal drafting and, perhaps, the English language. Use them thoughtfully and sparingly.

Series. Although authorities on punctuation may differ, when drafting Maine law or rules, don’t use a comma between the penultimate and the last item of a series. (p. 113)

Why use commas sparingly? To save on printer cartridges? And why thoughtfully?

I can think of two better ideas, especially for technical writing (and they’re more or less the same idea):

  • Use one rule for commas all the time and never think about commas again.
  • Use numbers or bullets and never think about commas again.

With all of this in mind, this is what the statute in this case actually looks like:

The overtime provision of this section does not apply to:

The canning, processing, preserving, freezing, drying, marketing, storing, packing for shipment or distribution of:
(1) Agricultural produce;
(2) Meat and fish products; and
(3) Perishable foods.

This language adheres to the Maine Legislative Drafting Manual. Outline lists are used, but not too extensively (it’s actually a little bizarre that the long list is inline and the short list is outline). The outline list (“Agricultural produce…”) has plenty of punctuation and a conjunction after the second to last item. Nothing is “flushed left”. And there’s no comma between “packing for shipment” and “or distribution” (whether there should be a comma is a question that’s worth $10 million to a Maine dairy company).

This is what the statute probably should look like:

The overtime provision of this section does not apply to the following items:

• Agricultural produce
• Meat products
• Fish products
• Perishable foods

This exception is limited to the following activities:

• Canning
• Processing
• Preserving
• Freezing
• Drying
• Marketing
• Storing
• Packing for shipment or distribution

No commas, no conjunctions, and no questions about what’s going on with packing, shipment, and distribution. It’s also probably a good idea to put the list of particular activities after the list of more general products/industries. It’s possible that the last item should be split in two (“Packing for shipment” and “Distribution” – although, like the appellant delivery drivers in this case, I would have expected the real statute to read “distributing” to match the other list items if this is actually how the statute is supposed to work). Who knows?

Whether you put a comma between the last two items in a serial list doesn’t even begin to scratch the surface of the problems with the drafting conventions observed in this statute.

Delayed Reaction: The 2016 LTN Tech Survey

December 15, 2016 —

Here are some observations on the 2016 LTN Tech Survey that came out in November.

Ratio of IT Staff to Users

About 84% of the 43 surveyed firms (36 firms) have one IT staffer for every 15–30 users.

The ratio of IT staff to users varies widely across businesses (Doron Gordon, CEO of Samanage, describes 1 to 70 as pretty standard), and it would be interesting to see how ratios at law firms compare to ratios at other professional service providers (accounting firms, consulting firms, investment banks, and so on). But one eye-opening point of comparison is IBM. According to an October 2016 presentation by Fletcher Previn, VP of Workplace as a Service, IBM has deployed about 90,000 Mac computers as of October, with 10,000 more projected to be deployed by December. These Mac computers are supported by five administrators. If we assume one Mac per user, this is an IT staff to user ratio of 1 to 20,000. (Between the small number of support staff and a number of other factors, IBM estimates that it’s saving $273–$543 per Mac over a four-year lifespan compared to a PC.)

Desktop vs Laptop

Twelve of the surveyed firms are planning to buy desktop computers during their next hardware refresh, and 31 are planning to buy laptop computers. All signs point to laptops being preferred over desktops for the last 10 years, so more law firms buying laptops instead of desktops is good news.

Limited Use of Cloud-Based Apps

Law firms seem to be using cloud-based apps mostly for “back-office” functions: data storage, HR, email, and accounting.

¹ Includes “Document management”, “Storage”, and the following items listed in “Other”: “Archiving”, “Disaster Recovery”, “DR”. ² Includes “Billing” and the following items listed in “Other”: “Accounts Payable”, “expense”, “Expense Reimbursement Processing”, “Time and Attendance; Expense Management”.

Only three apps were mentioned that seem like they’d be day-to-day performance enhancers for lawyers (all collaboration tools): Asana, ShareFile, and SmartSheet.

Reason for Limited Use of Cloud-Based Apps

Law firm system administrators seem to be really focused on data security (the first three items in this chart are arguably the same thing):

Wait… what?

The number of firms that (1) said they use cloud computing for “Storage” and (2) answered “No” to the only two specific questions about cloud storage (“Does the firm maintain its own back-up of business data that is stored in the cloud?” and “Do you use cloud-based storage and file-synchronization services like Dropbox and iCloud?”): 2.


Contracts Made Easy in Microsoft Word

December 12, 2016 —

Contracts are complicated, working with them is time-consuming, and it’s essential to make sure that everything in them is right. Contract Tools is a Microsoft Word add-in that makes creating and analyzing contracts simple. It’s easy to install, doesn’t require sharing documents over the internet, and fits seamlessly with attorneys’ working styles.

When you open a document in Word, Contract Tools analyzes it for things that often appear in contracts – provisions, defined terms, cross-references, and more – in almost no time at all: about a second for a 50-page document, two seconds for a 100-page document, and five seconds for a 200-page document. Contract Tools can analyze just about any kind of contract: short nondisclosure agreements, massive bond indentures, US contracts, UK contracts, Canadian contracts, and much more.

Once the analysis is finished, Contract Tools provides a variety of powerful features for working with contracts, including:

Proofreading: Contract Tools checks your document for common contract drafting problems: words and phrases that look like defined terms but aren’t, defined terms that are defined more than once, and more. Underlining makes drafting errors easy to see in context. You can easily ignore drafting errors that don’t bother you and restore ignored drafting errors if you change your mind.

To-Dos: The To-Dos view shows placeholders, bracketed text, highlighted text, comments, and similar items in an organized list. You can copy the contents of the To-Dos view to the clipboard with one click, so it’s easy to share a to-do list.

Defined Term Autocompletion: Most contracts have a few hundred defined terms, each used hundreds of times. With Contracts Tools, you can start typing a defined term to see a list of suggestions, and then choose the defined term you want to use.

Search: With search tokens, Contract Tools makes it easy to find numbers, amounts of money, dates, and other important information. You can mix tokens and text to form custom searches.

Contract Tools is the companion to Turner, a contract drafting application for Mac released in September 2015.

To find out more about Contract Tools, check out TechnoLawyer’s review of Contract Tools or get a free trial.

The ILTA/InsideLegal Technology Purchasing Survey: The Good, the Bad, and the Baffling

October 7, 2016 —

For the last 10 years or so, ILTA has partnered with InsideLegal to survey ILTA member law firms on their technology purchases. Here are some things we found interesting about this year’s survey compared to past surveys.

The Good: Data Storage and OCR

Law firms don’t seem to have as much interest in doing their own data storage as they used to, possibly because they’re outsourcing this to Dropbox, OneDrive, and so on. This is good news if only because law firm IT groups implementing Dropbox or OneDrive clones probably isn’t the best use of resources. (Interest in Microsoft SharePoint, which includes document storage features, has dwindled to almost nothing, possibly because of Microsoft OneDrive.)


(This chart and the ones below show the percentage of surveyed law firms who said ‘yes, we purchased something in this category in the last 12 months’ in a given survey.)

Also, law firms seem to have somewhat more interest in scanning and OCR than they used to (but this isn’t altogether clear – this chart could simply mean that all the scanners that law firms bought 10 years ago are finally being replaced). Anyway, well-constructed PDFs with searchable text data are better in almost every way than pieces of paper, so law firms being more interested in scanning and OCR is probably good news.


The Bad: Desktop PCs

Law firms’ continued interest in buying desktop PCs seems a bit backward.


With a few exceptions (that really only apply to non–law firm, third-party service providers – for example, sophisticated statistical analysis and CGI-heavy motion graphics for a courtroom demonstration), the work that lawyers do with computers doesn’t require desktop processing power.

Here are some data (from Gartner, as reported in a 2012 USA Today article) about desktop PC shipments compared to mobile PC shipments:


It looks like people have preferred mobile PCs to desktop PCs since 2007. I doubt this trend has changed since 2012. Mobile PCs just make the most sense for most people most of the time: they give you the tools you need to do your work whether you’re in your office, someone else’s office, offsite, on a plane, and so on. Why are 60% of law firms buying desktop PCs year after year?

The Baffling

“Soft Costs”

The surveys report law firms’ total technology expenditures as a percent of revenue (1%–4% is typical), but this amount excludes IT staffing and training. Why not include these items? The 2013 survey characterizes IT staffing and training as “soft costs,” but money spent on soft costs is still money, and it’s not as if these costs are set in stone. IBM’s recent experience with deploying Mac computers (as reported by CIO.com) makes for an interesting counterpoint:

IBM is already benefiting from the change of heart and organizational process, according to [Fletcher Previn, vice president of workplace-as-a-service at IBM]. The team of 24 IT staffers and specialists who support Macs at IBM is much smaller than what was required for PC support, and it spends less time fixing technical problems, Previn says. “You just have fewer problems coming in.”

While 40 percent of IBM’s PC users call the helpdesk for troubleshooting, on average only 5 percent of the company’s Mac user [sic] do the same, according to Previn. “The longer this program runs, the more compelling the business case becomes,” he says. “I can confidently say that every Mac that we buy is making and saving IBM money.”

Just how much money? “IBM tells us that each Mac is saving $270 compared to a traditional PC, thanks to much reduced support cost and better residual value,” said Luca Maestri, Apple CFO and senior vice president, during the company’s most recent earnings call.

No Crystal Ball

The surveys offer data on what law firms purchased in the last 12 months and what they plan to purchase in the next 12 months, broken down by a number of categories (for example, according to the 2016 survey, 61% of law firms purchased desktop hardware in the last 12 months, and 55% of law firms plan to purchase desktop hardware in the next 12 months). However, what law firms plan to purchase and what they actually purchase are often very different, and actual, historical purchases are a much better predictor of future law firm purchases than planned purchases.

For example: by category, 2016 purchases differed from 2015 purchases by 1.5% on average, with a standard deviation of 4.2% – in other words, 2016 purchases weren’t all that different from 2015 purchases. In contrast, 2016 purchases differed from 2015 planned purchases by 6.7% on average, with a standard deviation of 6.5% – a bigger difference.

This is similar to recent years:

  • 2015 purchases differed from 2014 purchases by −0.8% on average, with a standard deviation of 3.9%; 2015 purchases differed from 2014 planned purchases by 4.0% on average, with a standard deviation of 7.4%.
  • 2014 purchases differed from 2013 purchases by 2.1% on average, with a standard deviation of 5.1%; 2014 purchases differed from 2013 planned purchases by 8.3% on average, with a standard deviation of 8.5%.

It would be interesting to know what’s driving this: what makes technology purchases by category so difficult to predict, even when actual purchases by category don’t seem to change much from year to year?

Working with Contracts: Questionnaires and Wizards

August 17, 2016 —

We’re often asked whether Turner and Contract Tools are similar to LegalZoom and RocketLawyer. These are online tools for creating basic legal documents like wills and corporate charters. When you use either tool, you typically fill out a questionnaire, pay a nominal fee, and then get a finished document. For basic legal documents, these are good products for a good price – especially if you don’t need customized legal advice.

Turner and Contract Tools are very different from these kinds of tools. Here’s how.

Wizards: Questionnaires by Another Name

Turner and Contract Tools don’t use wizards.

More generally: software can present information in a variety of ways. Email applications present an overview of emails in a table, and details about a selected email in a larger view. Photo management applications present photos in an image gallery. Applications for creating music and film present media in a timeline.

Most applications for drafting contracts – not only online tools like LegalZoom and RocketLawyer, but also document assembly software like ContractExpress and HotDocs – present information with a wizard. In a wizard, a task is divided into a series of steps, and each step is presented one at a time – just like a questionnaire.

Wizards are useful for many kinds of tasks: buying plane tickets online, installing new software, filling out tax forms, and much more. But they’re not useful for everything.

Guidelines for Wizards

From experience using wizards in appropriate contexts (for example, an airline ticketing kiosk) and inappropriate contexts (just imagine getting started on an email by going step by tedious step through a wizard), we can sketch some guidelines for the kinds of tasks for which wizards are suitable.

Wizards are suitable for tasks that:

  • you don’t do often
  • you don’t fully understand
  • deal with information that you don’t care to explore
  • are inherently sequential

There are plenty of tasks with these traits. We all have limited time and attention, we can’t master everything, and we can’t fully understand and explore all the information we encounter every day. Wizards are a good solution for the kinds of tasks that we need to get done, but that we don’t need or want to fully understand.

Filling out tax forms is a good example of this kind of task:

  • Most people fill out tax forms once a year.
  • Tax forms (and tax law) are complex and arcane – few people fully understand them (or need to).
  • Unless you get audited, you generally don’t need to explore the information in your tax forms.
  • Filling out tax forms is inherently sequential (calculate income, apply deductions, and so on).

Wizards aren’t suitable for tasks that:

  • you do often
  • you already understand
  • deal with information that you may need to explore
  • can be done in a variety of ways

There are also plenty of tasks with these traits. For example, sending an email has each of these traits, so no one uses wizards to send an email.

Wizards and Contracts

For some people, drafting a contract is more like filling out tax forms than sending an email. If you don’t work with contracts often, you probably won’t mind (and would likely prefer) being led through a drafting process step by step.

For others – transactional lawyers in particular – drafting a contract is more like sending an email than filling out tax forms. Transactional lawyers work with contracts all the time. They understand contracts and contract law. They routinely need to explore the information in contracts and answer questions about what they mean. And they know that contracts can be put together in a variety of ways (lawyers have their own working styles, and working styles can change from transaction to transaction). For transactional lawyers (and anyone else who works with contracts often), wizards and questionnaires aren’t good ways to work with contracts.

Back to Turner and Contract Tools

There are many alternatives to wizards. Software can present information with an overview/detail framework, image galleries, timelines, and more.

How an application presents information is typically driven by the kind of information the application deals with. Photo management applications use image galleries because they’re a good way to present collections of photos. Music creation applications use timelines because they’re a good way to present time.

Contracts are complex, interconnected blocks of information, and both Turner and Contract Tools use a variety of approaches to help people deal with this complexity:

  • Extract information about connections from text and then present an overview.
  • Integrate information about connections into text (for example, by letting you explore connections by clicking words and phrases).
  • Check for broken and ambiguous connections, and then present this information in a brief menu or a more detailed list.

Interconnections in contracts are important, but contracts are also text, and text can get messy. Important information – payment amounts, deadlines, and so on – can be presented anywhere and in a variety of ways. Turner and Contract Tools offer tools to make it easy to find these less-structured kinds of information.

Also, contracts change: they’re negotiated, amended, and so on. Contract Tools and Turner (especially) offer a variety of tools for making changes to contracts: autocompleting text, tracking unfinished items, and more.

For transactional lawyers and others who work with contracts, the tools that Turner and Contract Tools offer are far better than a linear questionnaire.