Voice of the DBA

Category: Editorial

  • Fun with Dates and Times

    The other day I got a Github issue that asked if we here at SQLServerCentral should set our dates as YYYYMMDD in the new SQLServerCentral site. The issue noted that there were some inconsistent dates. Article dates are in a similar format, YYYY/MM/DD, but we do have some DD MMM YYYY and a few places where I think the American MMBBYYYY has lived on.

    In any case, I made a joke on Twitter of asking if people were alright moving to YYYYMMDD. Most people were in favor, though someone asked about MMMM DD, YYYY, which is easily read. I have to say that I somewhat like that for reading, but as data, as something I’d work with, I prefer having the year month day format. I do want separators, and while I grew up using slashes ( as in YYYY/MM/DD), I know the ISO format is dashes (YYYY-MM-DD). ISO 8601 has dashes, so I decided to add that to the GitHub issue as a way of formatting.

    What about using month names? As I look through various sites, I find inconsistencies. Some use dates, like Jan 11, 2019. Others might use 11 Jan 2019 or 2019-01-11. I find that MSDN uses regional settings, so the en-us version of a page will have 01/11/2019, but if I change to the en-Gb version, I see 11/01/2019. That seems find as a UI enhancement, but I hope that the actual value is stored as 2019-01-11. We don’t need them time

    Or do we? There are times involved. While most of us might not care about whether something was published at 8am or 9am, we might care about where it was published and the time. Certainly 8am in Christchurch if far different from 8am in Honolulu. Do we store time zones? Or keep all dates in one? The logical thing is to keep dates in UTC and then adjust based on the client settings to display something that matches their time zone. We’ve been storing data in UTC, though the server time has been the Mountain time zone in the US for historical reasons. I don’t like that dependency, so I’m sure we’ll change that.

    Dates and times are both simple and not simple. There are decisions to be made, and far too many of us take them lightly. While the display is of secondary importance, it’s likely that we’ll often use this data in other systems, as a way of marking a change, or even as a way of ordering our data. Having consistent and known date time formats is important as a data professional. I hope you pay attention to how you gather and store this data.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 4.3MB) podcast or subscribe to the feed at iTunes and Libsyn.

  • Measuring Effectiveness

    I’ve never been the guy that produced code the fastest, or the most code in a day. I haven’t been able to show that I work the most hours, or that I even keep great hours. I’m often late or erratic in the times I come and go from work. I don’t show that I am closing tickets in consistent times as I often will be slow to close new issues. I haven’t ever been the employee of the month.

    What I have been is effective. I get things done. I leave the organization and the job better than when I found it. I work to deliver value for the salary I’m paid. That isn’t something you can easily measure with many traditional metrics that people use, but there are ways to decide that I’m a good employee. It’s a little ephemeral and sometimes uncomfortable to have an employee that doesn’t seem to fit the same model as most other employees, but that doesn’t mean you wouldn’t want me working on your staff.

    This article on measuring DevOps tries to explain how you can determine if your DevOps process is working, which is a similar ephemeral way of working. When we implement DevOps, we don’t have large project plans, we don’t look forward to completing a system. We undertake a set of work, knowing that we don’t have an end date. We just keep doing work and getting things done that are needed. In this case, how can we measure if a project was on time, on budget, and finished?

    I’d argue, as would the article, that those aren’t good measurements. Those are the ways we would traditionally look at work, but in software, those aren’t often the things we want to look at. If we deliver software that meets some requirements, but users struggle to use it or complain, is it done? Most management would call it done, just like the airlines mark the plane as leaving on time when it pushes back from the gate. Standing on the runway for 45 minutes doesn’t count against that target, even though I’d call that leaving late.

    The article talks about picking things that bother a manager or team members about the software or their process. Some of these are easy, like the time to deliver a new server or recover from a failure. Others that I’ve found to be useful are the time to assemble a release or deploy software to an environment. While we can sometimes play with what these numbers measure by moving to smaller set of changes or altering our process, we’ll quickly find out if we have a lot of overhead in our work by adding these measurements.

    What about the software itself? Can I actually compare the time to deliver report a vs. report b? One might be much more complex than the other. That’s true, but we ought to average out delivery estimates over time, and we can certainly separate out query writing of the complex logic against the time required for formatting. In fact, it might be good to start to measure different parts of software delivery to find out if certain people are better at some parts, or if requirements from clients are causing unnecessary delays. I know I’ve certainly had some formatting complaints require more time to get right than the entire rest of the software.

    It can be hard to develop metrics that have outcomes and are truly actionable, but it’s a better way to determine if your team is improving. Relying on simple and traditional metrics is lazy, and allows for lots of argument and debate over whether a team is doing well or not. Tackling those specific items that are irritating to management or customers might take some work, but you’ll end up with a list of things that can be targeted for improvement and show progress that actually means something.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 4.8MB) podcast or subscribe to the feed at iTunes and Libsyn.

  • Badly Trained AI

    Most of us know that data is being used to make more and more decisions inside of all kinds of organizations from retail giants to banks to sports teams. We are constantly asked, or see reports, of data driven decisions. We often need to show some data that supports and explains the rationale for making some choice. As our populace becomes more data savvy, I expect this trend to continue.

    AI (Artificial Intelligence), and the related Machine Learning (ML), are becoming more and more widely used. From mobile phones to autos to trading systems, we regularly see new “AI capabilities” being added to products and services. No business or industry seems immune, and I’m sure many of you are seeing AI being incorporated or feeling pressure to start using some AI in your work. As you work with AI, or start to, you’ll quickly realize the importance of data in your efforts.

    This is true for the cleanliness of data, but perhaps even more important in the tagging of data sets. As Amazon learned, building an AI or ML system, is hard. They scrapped one system that was being used to rate resumes and help their recruiters sort through the volume of applications they received. Why? Because of bias.

    Apparently the system would downgrade women’s resumes for various reasons. To me, this is a perfect example of a principle I’ve had throughout my career: garbage in garbage out. In this case it’s not necessarily bad data that was the problem, but bad tagging of what was a good and bad resume, probably from the internal prejudices of a few people.

    There will be more dangers as we use ML and AI technologies in our work. It won’t be enough that we clean the raw data for training, but also that we clean and properly manage the tagging of what data sets represent the results we are looking for. Like in much of our software, it’s easy for us to only consider the happy path, to only tag those items we think are good results. That is useful, but we might also be unconsciously tagging other results as bad, which appears to have happened to Amazon.

    We can build systems that do a better, more rational job than most humans, but we need extraordinary care to ensure our training data lacks bias. Unfortunately, most people both think they’re not biased and are unwilling to spend extra resources to deeply examine the data. Two things that worry me about the future of our AI/ML systems that will inform us.

    Steve Jones

  • Republish: MY Data

    As I noted, I’ve been ill. This is being republished to give me another day to get well: MY Data.