Friday, November 17, 2017

The Eisenhower Matrix Fallacy

Road trips reveal truths and clarify ideas, at least that's what my Auntie Pat Tern insists. On a recent trip with her and cousin Tim Toady,  this came up when Tim requested a bathroom stop. What he said was, "I could use a bathroom some time soon." And as we flew past exit after exit, Auntie Pat Tern said he should let her know if it becomes urgent, otherwise she wanted to find a good spot to stop for lunch too. That got me thinking about competing interests when it comes to programming and the Fallacy of the Eisenhower Matrix.

Jastrow, J. (1899). Popular Science Monthly US public domain
Duck or Rabbit? It may depend on how long

you allow yourself to analyze the problem.
In a speech at Northwestern University, President Dwight D. Eisenhower said "I have two kinds of problems, the urgent and the important. The urgent are not important, and the important are never urgent. "   


Prioritizing Competing Requirements

Important issues and urgent ones often compete. On our road trip, Auntie Pat Tern knew that a nice place to stop for lunch would also have a nice bathroom, that was part of her selection criteria. She also wanted to get as far as possible to her ultimate goal before the end of the day. Stopping both for lunch and for a separate bathroom pit-stop would delay her from reaching her ultimate goal. So she employed the Eisenhower statement as it was intended, to express that a thing cannot be both urgent and important.

Unfortunately, a lot of people in the business world have worked Eisenhower's statement into a matrix, which suggests that requirements can be both urgent and important.  Trying to accomplish a task as both important and urgent makes no sense. If it is important enough to need an optimal solution, then take the time to discover that optimal solution. If it is urgent enough to resolve immediately, do your best with what you have right now.

Importance means you have the time to be thoughtful and take time to discover the most optimal solution; it is the polar opposite of urgency.

As long as stopping for a bathroom break was important, it was worth waiting for the best location to stop both for food, which was also important, and for clean restrooms. But as soon as the need for a bathroom became urgent, the competing priority for food could be disregarded, at least to some extent. And with greater urgency, even the competing priority for a clean restroom might be set aside; the greatest urgency means you act now even if you have to pull off the road at that tree up ahead.

Eisenhower and Software Architecture 

For software architecture, the urgent and the important need to be kept separate because the important parts of the project should inevitably be the most stable code in the system. The least stable code is the stuff you write out of urgency, when things are likely to change, but we can't wait until we understand all the factors that cause change.

Don't believe the Eisenhower Matrix, it is a fallacy. Important and Urgent are like North and South ends of a magnet. Anything that is equally Important and Urgent lies in the middle, and that part of the magnet is neutral. The neutral zone is where we find the mindless and repetitive tasks that only become urgent when they are neglected for too long and only become important when something out of the ordinary arises.

For Tim Toady on our road trip, he appreciated waiting for a sandwich on Dutch Crunch and a clean restroom because he values good food; it's important as long as he doesn't have more urgent needs.

Sunday, November 12, 2017

Why Coders Use Interfaces

Going on a road trip with Auntie Pat Tern and cousin Tim Toady reveals a lot about them as only a road trip can. Tim was called a "whipper snapper" on more than one occasion for criticizing Auntie Pat Tern's driving. "I cannot wait for self-driving cars," he proclaimed while Auntie Pat Tern praised the freedom of the open road and the control she feels while driving a car. Which got me thinking about why interfaces are so important as part of my programming team's design decisions.

Tim Toady prefers to drive arcade games while Auntie Pat Tern craves a 64 1/2 mustang convertible. But they both drive reasonably well in the electric car they have at home and the gas guzzler we rented for our long drive. That's because the interface of gear shift, accelerator, brake, and steering wheel are universally well understood by users.

Accelerate, brake, steer, and shift.
The simplicity of a good interface.
That same interface translates inputs to a multitude of different systems from arcade games to tractor-trailer trucks. When it comes to deciding whether to accelerate or decelerate, the driver/user does not care if the vehicle is electric or gas-powered. A lot of people who drive electric cars even still say "gas pedal".

Autonomous driving vehicles use the same interface as people: shift, accelerate, brake, steer. By using the same interface, a person can take over in an emergency, or maybe just for the fun of driving.

Developers use good interfaces like this to future-proof their code. If the business rules are protected from the mechanism for inputting data thanks to a reliable interface, we can make sure the most important business rules continue to function as expected.

Sure, sometimes the interface has to change a bit with new business automations. When cars got automatic transmissions, the gear shift changed and we lost the clutch. The same happens with code, new automations require small changes to the programming interface, but the goal of an interface is to minimize the need for such changes. Ultimately, we want to put interfaces between systems to give ourselves flexibility to support new automation, like Tim Toady's self driving car, as easily as we support more manual systems, like Auntie Pat Tern's stick shift.

Wednesday, April 5, 2017

Real Business Data Returns

Auntie Pat Tern believes you can have your cake and eat it too, "Just use your resources wisely". So I wanted to see how to use my static resources more wisely in my Salesforce org. I have previously shown how to use static resources for test data. Now, I want to use that same data in developer sandboxes created from my org.  Since static resources are brought over like code to developer sandboxes, I can reference them in code I execute when I generate the sandboxes. Using the map datatype helps me keep relationships intact as well:

With this simple trick, I can preload my dev orgs with sample business data that I am using in my unit tests. Like Auntie Pat Tern promised, I can have my cake and eat it too when it comes to using realistic sample data in my tests and development orgs.

Whole Heaps of Fun

Auntie Pat Tern has noticed that questions about heap limits come up periodically. This could be a concern for developers needing to work with files and attachments in Salesforce. Luckily, those native objects aren't subject to typical heap limits. If you stick to blobs you can work with these large data objects. Don't use this knowledge to do bad things, now! Limits help us write code that performs well for our users as well as our fellow tenants in the multi-tenant environment we all know and love.

12:52:35:797 USER_DEBUG [43]|DEBUG|Heap: 57,789,821 / 6,000,000

Of course, if you test your code at limits, you can be sure of just how awesome your code is, which is another great reason to write thorough tests for all your code. You will find that if you are careful of what datatype you use and what you try to do with it the platform rewards you with heaps and heaps of fun.

Wednesday, September 28, 2016

Eliminate Cargo Cult Programming and Build a Stronger Team

My Auntie Pat Tern is one of those people who refuses to let family members wear cargo pants when we go out to dinner.  She says she doesn't want to participate in the 'cargo cult'.  She inspired me to look at how Cargo Cult Programming has affected our Salesforce org.

One of the best things about programming on the Salesforce platform is the wealth of resources online to help problem solve. It can be one of the worst things about programming on the platform as well if you aren't careful about the resources you choose to follow.  Some code samples use outdated practices, others may be snippets of code taken out of context, which may not behave as expected in a different context.

In one case, the Apex documentation team included System.assert statements in a code snippet.  Our new developer later copied the snippet into a utility class.  The developer found a great resource but didn't fully understand the difference between System.assert and System.debug.  The former builds in a fatal error, appropriate for test code but not for our essential platform automations. Sniffing out this cargo cult programming helped us know what additional training could help our developers.

A more frequently found cargo cult is the factory class to build basic objects. A factory class is still far better than using the very outdated SeeAllData in your test code.  But test data factory classes add unnecessary work for coders. Salesforce offers a newer and better technique with Test.loadData, which I wrote about in July.  It lets us shift the work of creating and maintaining sample data from the developer to the administrator.  

Other cargo cult programming that pops up in our Salesforce org relates to data sharing and security rules.  We teach our developers early on to understand the difference between code that runs only in Triggers vs code that may be run by end-users via UI customizations like pages, flows and actions. And we encourage our administrators to monitor org security in our code during code reviews and test executions.

Hunting out and eliminating cargo cult programming helps us build cohesion among the team that maintains our Salesforce platform and better understanding between administrators and developers who now work more closely together. This helps us spread the effort of org maintenance across a team and make use of clicks and code for our projects. End users benefit and our org is more reliable and easier to maintain as well.

Friday, September 9, 2016

So Many Resources, So Little Time

I found Auntie Pat Tern burying a hammer in the flower bed. When I asked why, she said it was Tim Toady's favorite tool and he needed to learn something new. When all you have is a hammer, everything looks like a nail, so she wanted to make sure he found new ways of doing things. I like to make sure my developers are constantly learning new skills as well, so they don't rely on "golden hammer" practices that may have become outdated.
Become a specialist and then learn even more.

Learning to develop on the Salesforce platform is like finding a magpie's nest full of shiny objects. Where do you start and how do you figure out which gems are most valuable? Here's the path I take:
  1. Attend Dreamforce. The keynote and product demos will help you know what you want to learn. The sessions will help you get started on that learning. It offers the chance to talk to people who are on the same learning journey as you are and share tips and interests. I am presenting two sessions on developing with Apex this year: 8 Essential Apex Tips for Admins and Apex Trigger Essentials for Admins.
  2. Watch more Dreamforce sessions online. After the event, Salesforce makes sessions available by video online. Sessions you wish you had attended, topics you didn't know you were interested in until after Dreamforce, all are free to watch online.
  3. Follow up your learning in the Success Community.  Find your local user group and developer group and follow the online conversations to learn how other people are using Salesforce and the challenges they are overcoming. Even if you don't have a local group, you can join groups online to be part of the discussions.
  4. Keep up the learning with Trailhead. Salesforce offers online learning opportunities through Trailhead. Don't be intimidated by the SuperBadges, they offer a well defined path to learning about a particular skill in Salesforce. And if you get stuck the community is there to help.
  5. Two places to turn for online help from the community when you get stuck on any of your coding projects are the Developer Community and Stack Exchange. Search the boards to see if someone has already asked about the question you have, and if you can't find a discussion, go ahead and post your question. Once you gain more skills, you should find that you are answering more questions than you are asking.
The resources are there to help you along. Learning to develop, or learning to develop even better, can be fun with all of the resources Salesforce offers. You don't have to rely on golden hammers when there are opportunities to learn new and better ways to work with Salesforce.

Sunday, August 7, 2016

CSV Data: Commonly Surfacing Vexations in Data

According to Auntie Pat Tern, "Trying the same thing over and over without getting the results you want is enough to make anyone crazy." Unfortunately, that's the excuse Tim Toady uses when he doesn't do his homework because he's tired of not getting the results he wants from the effort.

When it comes to creating CSV files for use as unit test data, what should be an easy process can make you a bit crazy if you wind up getting unfamiliar errors. The following steps and potential errors may help:

Step 1: Export some production data.
Step 2: Delete all but a reasonable selection of data.
Step 3: Remove system fields.
Step 4: Make sure the date fields are in the form YYYY-MM-DD
Step 5: Make sure date/time fields have a "T" between the date and time, with no space, such as 2016-08-07T01:40:39
Step 6: Make sure the commas that appear in text fields haven't thrown off the number of columns.
Step 7: Make sure there are no blank rows with only commas and no values
Step 8: Renumber IDs starting with 1
Step 9: Remove any invalid IDs for related records (Account ID of 000000000000000AAA appears for all top-level accounts, and should be removed, for example)
Step 10: Upload the CSV file as a Static Resource for your code to reference

If you are creating a set of child records for an object that is related to another object, you can sort the parent and child data by the IDs of the parent to make sure you get records that match in both data sets. Use search and replace to redo the parent IDs to match the new IDs in the CSV file of parent records. For example, if you have a CSV for Accounts with IDs numbered 1-200, the related contact records must use 1-200 for the Account ID as well.

Bad CSV files might result in the following errors:

Potential Error Likely Solution
Invalid ID value on line 2: 000000000000000 Remove invalid IDs
Too many DML rows: 10001 Load fewer records
CSV Parse error: '8/20/1959' is not a valid value for the type xsd:date Format dates as YYYY-MM-DD
Duplicate ID value on line 81: null Remove empty rows from CSV file
CSV Parse error: '2011-09-07 01:00:31' is not a valid value for the type xsd:dateTime Format date/time fields with "T" rather than space between date and time
Validation Errors While Saving Record(s) Erroneous data or IDs not starting with 1
System.UnexpectedException: Salesforce System Error Remove stray commas throwing off columns
Static Resource not found Make sure code refers to Static Resource by name

Start with a small number of records using the fewest fields for testing the code or configuration changes you need to test. That way, you won't wind up like Tim Toady, who falls back on bad habits when errors occur with his first attempts to follow best practices.