Snips

2026

January. I am cutting back significantly on posting content across social media. However, I also don't want to re-write some of the below posts from the past few years, so I'll leave these linked below this.

• Databricks and meta-information: https://x.com/sqlinsix/status/1774756248736391425

• Creating a date dimension in python pandas: https://sqlinsix.medium.com/how-to-create-a-date-dimension-in-python-pandas-5e61bce72256

• Using a broadcast over a shuffle: https://x.com/sqlinsix/status/1825479846039785833

• Triggers and ADF: https://x.com/sqlinsix/status/1881326259587764548

February. Unlikely - https://finance.yahoo.com/news/bank-america-revamps-silver-stock-030300865.html. In my view, I'd be surprised if a troy ounce of silver was above $100 by the end of 2026. Now oil..

August. Estimated total shares outstanding multiplied by average price last twelve months is $20,864,000 with a net income of a little over $3,200,000 for the past four quarters (6.5 P/E). That's an earnings yield of 15.38%. And they say value investing is dead!

October. While a contrarian take for now, LLMs are like a trust fund kid giving a lecture to a holocaust survivor on the value of perseverance.

2025

January. Notice the books you can buy versus the ones you can't anymore. What's the pattern?

May. As of this year, a grain I track costs $0.622 per ounce. In 2009 it cost $0.087 per ounce.

June. Out of pocket cost for urgent care visit in 2000: $15. Out of pocket cost for urgent care visit in 2024: $375. Only the provider and office differed.

July 1. Udemy students in the Automating ETL course - see recent announcements on that page.

July 16. The happiest men and women late in life? Grandparents.

August. I had the same reaction as Jeremy at 5:07 - https://www.youtube.com/watch?v=fM1gSOCNV0E&lc=.

October. If you get significant views relative to friends, connections, subscribers or followers (FCSF), you may have contrarian thoughts. Or, you may be resold (easier now). Few consistently get five multiples views relative to their FCSF. Reduce sharing anything and procure.

December 3. Never trust someone who needs and seeks attention. The statistical reality of that person is they'll engage in left tail behavior as their experiences regress to the mean.

December 20. Will 16:02 foreshadow the future - https://www.youtube.com/watch?v=d3R5MZGGjzU?

2024

February. Tom - "Know the client needs even better than they know them." Example - "And you also want [x]" with the client like, "Oh yeah, we do want that." The guy is phenomenal at consulting.

April. SqlinSix contributed some research to this article about AI and the future of work.

September. This assumes that knowledge and experience are the same - https://x.com/austin_rief/status/1835418133110374686.

November. Things don't matter until they do.

2023

February. In Poor Charlie's Almanack, Charlie Munger cautions about the self-regard tendency. Required reading for consultants in my view.

March. A+ advertisement in my view - https://www.youtube.com/watch?v=GrnkY4buucM.

August. Martha - "You know how to make the best of a bad situation." Best neg I've heard in a while.

September. A physical copy of The Art of Worldly Wisdom dropped from $11.99 to $2.. and they say value investing is dead!

2022

March. Publishing the last of my YouTube videos early this year - https://www.youtube.com/watch?v=DtXoXAOjjB8. After almost ten years of posting videos on YouTube, I am saying goodbye. It's been fun, but our society is no longer protecting private property with information. One counterpoint: this is cheap prices curing cheap prices (information).

December. I mostly agree with Charlie Munger that AI is the new ***. Many AI applications will not offset their costs in the long run. Expect it to go through a big hype cycle before people recognize this.

2021

January. Udemy students in the Automating ETL course - last year, I added several videos that relate to understanding basic data principles. I plan to keep most of these added lectures for at least a year, but a few of these videos are time-relevant so they will less relevant in a few years.

April. Ignorance as the actual security technique, even if some don't realize it - https://www.youtube.com/watch?v=LHZmg48sD7Y. Unrelated to that video, but fundamentally all crypto-tokens belong to the "you will own nothing and be happy" since all are fundamentally nothing. It's peculiar that no one in the community realizes that. But in my view, bitcoinaires are the kind of people who know the price of everything and the value of nothing.

May. Kitco interviewed Nouriel Roubini - https://www.youtube.com/watch?v=ZDq3hMoZ8XE. That is a signal in and of itself. Pretty soon Kenneth Rogoff may be on Kitco!

November. Worth remembering: spark dataframes are computations.

December. One of my mentors, Matt, suggested that I read Psycho-Cybernetics (Dr. Maxwell Maltz). I am so thankful I didn't read this book in my 20s.

2020

[I added quite a bit of content for Automating ETL students in the Udemy version.]

June. I saw a sign outside a well-known store that read, "We buy houses, cars, trucks, land, anything!!!" Does that guy know something?

2019

January. Preventing SQL injection - https://www.mssqltips.com/sqlservertip/5856/validate-dynamic-sql-to-prevent-sql-injection-in-sql-server/.

Remember that the best data practice for companies: don't store data you don't want compromised. Attackers inherently don't follow rules, thus they have an advantage.

March. The past year, I created a simple script that scanned through proposed T-SQL changes. It catches sneaky development.

Two major takeaways that I learned from this script.

First, it doesn't use any unnecessary library and because I use base libraries like StreamReader and StreamWriter, I also stay away from wrapper libraries. I have noted throughout my career how rare I've seen the base libraries used because it requires a little more critical thought. However, they carry advantages because I can keep the functionality simple and reduce security compromises from complexity.

Once we started catching some of the sneaky development, we found other patterns. Because of how I designed the functionality (see above point), this made it easy to scale. The script is less than 80 lines total, yet does more than most of the other libraries I've seen out there, plus has far fewer libraries and complexity.

May. Using PowerShell to encode files - https://www.mssqltips.com/sqlservertip/5966/using-powershell-to-encode-files/. I've been using a lot more Chinese Mandarin recently, so this was something I've been playing around with testing.

June. The cloud feels cheap, so this article won't be an appealing solution for many firms right now - https://www.sqlshack.com/finding-unused-resources-impacting-azure-costs/. Over time, I expect this to be more of an issue.

This can all be prevented if the cloud environment is both planned and organized well.

July. Chitra - "We ask devops for help with everything."

I've noticed.

September. The theme this year, "If at first you don't succeed, you can dust it off and ask devops."

October. A tornado interrupting *** might be the most Texas thing that has ever happened to me. At least I was in a safer building than the last time a tornado struck.

December 4. Tiffany - "We have to do that or else we're considered [something else]."

To Tiffany's credit, she acknowledged that this rule made no sense.

This data point aligns with some other data points and paints a clear picture

December 19. Think of your filters by how you would horizontally scale. Source system? Jurisdiction? Year (date)? From there, running code in parallel becomes much easier.

This assumes that you've done a bit of assessment on how balanced vs skewed this particular filter would be. Obviously if one category has one-thousand multiples the others then you end up with uneven scaling.

2018

January 2. Why ceteris paribus matters in statistics - https://www.youtube.com/watch?v=RTkghV2fsoc. The inverse of what I share in this video is one sneaky way to lie to people with statistics.

"Sarah makes more money than Mary!" Reality: Sarah has 10 years of experience while Mary has 1 year of experience.

January 24. Data tendency and spread (examples) - https://www.youtube.com/watch?v=zpnC-aH9oSQ. The example at 6:05 is one of my favorites, plus makes a great data point that explains a lot of sociocultural patterns.

January 31. I'm noticing a pattern with both companies and clients I've worked alongside. They hire costly developers who recommend spending millions on software tools. Then these software tools require a lot of time to learn, so the company then spends million training their team on these software tools. Later these software tools make changes that require further learning plus will often change the terms and conditions that cost more money for the companies using the tools.

Why would I hire a software developer to recommend a tool?

No dude, your job is to connect your skill (creating software) to solving our company's problems (tool) while protecting our intellectual property (data).

That last point is also key: these companies are actually giving away their intellectual property by using a lot of these software tools.

The following is a list of software tools I've actually recommended using in my career: .

February. Yes, the cloud makes it easy to build a hundred environments for isolating specific needs. But how many specific needs do you actually have that you cannot consolidate in your testing? Will you feel proud of this spend later?

March 3. In traveling across Asia over the years, they view gold very differently than the West. For instance, the West assumes gold caused the great depression. Some teach this in Western institutions, though I'm not sure when people came to that conclusion.

In speaking with some people in China, they view their silver standard as almost preventing them from entering the great depression. Until their central bank started to sell silver to the West (actually sell it), they didn't struggle. However, shrinking the money supply (selling silver) in their mind is what caused their problems.

I heard similar themes in Singapore, India, Vietnam and other places in terms of favorability toward gold. I don't think I heard a single person describe gold as a "barbaric relic", which I've heard at least twice here in the US.

This perception affects the production versus consumption mindset. For instance, the West has GDP (C+I+G+NE). GDP seems backwards to a lot of people in Asian countries. In listening to them talk, they view GDP as being a measure manipulatable on the consumptive (demand) side. Governments could create a lot of money and spend it - boosting GDP. But more spending doesn't automatically mean a higher standard of living, while increased production of a good that's needed could result in lower prices (or more competitive pricing). Also, many of them pointed out that more production over time led to more innovation, as people get better at producing when they produce. I hadn't thought of that, but it's a strong argument in favor of manufacturing in general.

To be fair to the economist who thought of GDP, he thought of it during the gold standard (1934). Thus every unit of consumption represented a fixed standard of production; for instance, a $1 spent equated to 1/35th a troy ounce of gold. You cannot say the same when $1 is something that can be printed or digitally created but is not fixed to any value.

March 26. Jayson shared today about some of his cybersecurity observations while working at institutions that needed to protect wire rooms. He conceded a point that I remember from Zero Days: air gapped networks don't exist. He shared some fascinating stories from his experience and chuckled at people who talked about air gaps like they were secure.

One article he shared that I've never heard of before came from Wired (https://www.wired.com/2016/06/clever-attack-uses-sound-computers-fan-steal-data/). This attack involved using a computer's fan to extract data.

I looked up the researcher and he's on YouTube with some examples of cybersecurity compromises - https://www.youtube.com/@cybersecuritylabsbengurion7859/videos. Scary stuff and it goes to show that the things you don't know can have an impact.

I keep saying: digital security is an oxymoron.

I created a security page for me to track information related to security over time - https://sqlinsix.com/research-2018-security-page.

May 5. Peter Thiel's observation about how competition affects a person's identity is very similar to what I observed with how attention affects a person's identity - https://www.youtube.com/watch?v=3Fx5Q8xGU8k. A person wins the competition, but loses their identity. Or, as I once observed, a person gets attention, but loses their identity in the process. Fantastic lecture worth reviewing from time-to-time, as he drops a lot of gold.

May 16. I'm surprised how many companies miss the importance of senior staff. I overhead a company lose an $85 million contract because they were not able to deliver on time. The company eliminated their senior staff to save $700K per year.

The good news is that in 121 years, that company will recover that lost contract through savings.

June 5. Tracking real-time latency issues in SQL Server - https://www.mssqltips.com/sqlservertip/5398/solutions-to-track-delayed-sql-server-transaction-latency/.

June 13. No consistency errors. Split point LSN suddenly. No size increase. Integrity check requires almost 3 hours to run versus fewer than 5 minutes.

Great reminder that I can't always assume that "no errors" means no errors.

July. Using the lowest unit of measurement - https://www.sqlshack.com/sql-server-using-lowest-unit-of-measurement-in-t-sql/.

Don't confuse this with "the most accurate numeraire," though in some cases, the lowest unit can be the most accurate numeraire.

August. Udemy students in the Automating ETL course - I've added an announcement recently about a few best practices when developing with new technology. Those make good points to consider as a starting point. Remember to defer to your own experience.

September. One interesting observation I've seen with all of James Rickard's books: when you think about them from the frame of cybersecurity, they make sense on a much different level. When I listen to most people discuss his books, they mostly discuss his book from a financial view (or a global monetary view). But when I read his books from a cybersecurity perspective, I come away with a slightly different perspective than the simple monetary summaries I hear (or read) from others.

Maybe I missed the part where he comes out overtly and says it, but cybersecurity seems to be an underlying theme in his books. For example, if you consider money from a cybersecurity view point, then think about "physical gold" in that context (security), you come away with a much different view than if you think about "physical gold" only in terms of finance.

With digital money, one thing to consider: I may get a 10,000% return in my digital fantasy, but after the hackers get it, my return was actually 0%. That doesn't apply to the physical world (though Mother Nature is her own hacker of sorts when it comes to life).

What gold people do miss when they mention gold being secure due to its physical nature is that this also can involve other physical resources.

December 3. No data will always be better than bad data.

December 31. My close friend Steve and I ended our crypto media project. I am blessed to have learned quite a bit about media. This entire project answered patterns I've seen in journalism that I never understood. I have a lot more respect for media now because media production takes work.

Steve and I also co-wrote a book together, The Little Book of Positive Returns In Negative Rates. We enjoyed sharing areas of experience that we understood and co-writing a book makes this easier.

2017

January. You know who likes the Phoenix Project (Gene Kim, Kevin Behr, George Spafford, Chris Ruen) more than devops? Hackers.

February. I watched the documentary Zero Days (by Alex Gibney) recently. I had to watch it three more times because I learned a ton plus there's so many details that were easy to overlook. Talk about a book of notes! Something as simple as words versus pictures can have a huge impact on security and yet how many of us consider that?

I highly recommend this documentary.

May. James announced that he wants Fred to help with the Dallas Ethereum community. Congrats to Fred and great addition. Fred loves learning more about the cryptosphere.

June. Security continues to be a top concern and I expect this trend to continue. I'm surprised how many security professionals I meet do not know what sim-swapping is. Scary! Not related to sim-swaps but SQL Server instead, here's a recently article about audits for SQL Server - https://www.sqlshack.com/sql-server-database-architecture-and-audits-for-enhanced-security/.

July. I went through their entire certified ethical hacking video series on Pluralsight. Most of the videos were taught by Dale Meredith - great guy. He repeatedly said that "You can't stop hackers.. at best, you can delay them."

I highly recommend Dale's videos. Phenomenal teacher!

August 4. John - "You approach problems like a chess player. I've noticed that you quickly tend to be aware of a solution's costs, while everyone else is focusing on how the solution benefits them and later realizing the costs."

Don't all risk-averse people think this way?

August 21. I did not understand Durango. I did enjoy everything else.

November. I just finished Cybersecurity and Cyberwar by P. W. Singer and Allan Friedman. After reading it, this continues to confirm my thought that digital security is an oxymoron. Like Zero Days, this book confirms that the best approach to private data is to avoid storing it.

This book also has some shocking stories.

December. I cover some benefits to using clustered columnstore indexes in SQL Server for ETL applications - https://www.mssqltips.com/sqlservertip/5225/sql-server-clustered-columnstore-index-examples-for-etl/.

2016

January. One pattern I've seen with DBAs who start their career managing one or two servers versus DBAs who manage hundreds: the latter DBA group tend to think in scale from the beginning. I watched a DBA solve a problem a few months back and noticed how manual he approached his problem. It makes sense because he manages one server. However, his solution wouldn't scale if he had to manage one hundred servers.

This may partially explain why I see some people use the tools they use. They can use those tools because they're only solving one or two problems. It doesn't scale, but their situation doesn't require thinking in scale.

February. Vivek spoke at our discussion last week about clustered columnstore indexes. I did a little playing around with them this weekend and do see some big advantages with tables that involve heavy aggregates or where I would need data compression. For instance, I seldom access archived data, so this would be extremely useful for those data.

One might say, "But if you need a particular data point, then it would take longer than if you used a b-tree index." This is true, but if I'm only getting that data point once or twice a year, that longer cost is no big deal whereas a higher storage cost might be more painful.

Short video on some of my experiments - https://www.youtube.com/watch?v=mtmECeWrAEo.

March 10. I love Rob McEwen; he reminds me of my granny - https://www.youtube.com/watch?v=_y5SLyuYGtU. I seldom see this point raised about mining. That's peculiar.

March 27. Documentation best practices - https://www.youtube.com/watch?v=LQ2fFJhNWVc. Writing forces both critical thought and organization.

April 28. MSsubscriptions in Status 3 - https://www.youtube.com/watch?v=flCvOeCQp88.

May 3. In my view, this article written by Kenneth Rogoff will age extremely well https://www.project-syndicate.org/commentary/gold-as-emerging-market-reserve-asset-by-kenneth-rogoff-2016-05. I'll definitely keep his comparison of gold and short term treasury yields in mind going forward.

May 16. I actually like SQL Server replication. I do think many people don't think through how they set it up or approach it from a monitoring perspective, then this costs them. This explains why many people write about replication from a negative viewpoint - they're not critically thinking about what's involved.

Either way, here's a recent article I wrote for the SQLShack community - https://www.sqlshack.com/some-replication-architecture-errors-and-their-resolutions/.

July. In my view, local gyms make the best tourist souvenir.

November. My mother told me today that one of her brothers shared that uranium industry was cheap (mining, processing and the actual element). The last time I heard, her brother works as an accountant for a healthcare firm.

December. While ride-sharing today, I spoke with John, a part-time rancher. Cool guy. He confirmed what I learned on the farm about good mother animals - they are your prize. A good mother can produce many multiples of themselves (same with goats and chickens). You keep those prizes as long as you can.

2015

January. You can save a lot of time by noting what people recommend or suggest. If you're starting at 2 and someone started at 21, then you won't connect the dots because you haven't arrived at 21, much less 20. A 2 will best understand a 3 and may understand a 4, 5 or 6.

March. Adrian shared with me today that in his observation, object oriented developers approach databases as a trashcan. He's right in that they often don't think of how they'll get the data. This is one reason that object-oriented developers tend to dislike the SQL language because the SQL language challenges you to think of data sets, not objects. Anyone can shove data into a database, but when you need to get that data, you have to think about how you'll get that data relative to all the other data that you have.

How does one data point differ from the entire set of data? Or, if you need to return a set of the entire set of data, how would you identify that?

I remember a while back, Chris telling me that he understood where MongoDB could be useful, but he felt that most object oriented developers preferred it because they hated thinking about data storage. Granted, Chris is an exceptional C++ developer who understands SQL well (no surprise). I think MongoDB has a place, but I agree that MongoDB and any database approached with a trashcan approach will end up causing suffering for everyone.

Databases are organizational structures. In the same way that we organize our desk (how do I quickly get what I need to use), we should organize how we store our data.

June. Sheldon is right about the star-rating versus like-rating - https://www.youtube.com/watch?v=E94l1KGC24E.

The "data" reason for why: star-ratings cost more than like-ratings. Also, applied data to marketing shows that in this context like-ratings get more people to consume the content over star-ratings. The goal of some platforms is to keep people on them as long as possible, not get people to find the best content. A star-rating helps with getting people to find the best content, but can significantly reduce time on the platform.

July. With Texas continued growth, the state must address its drought problem sooner rather than later. One of my Aunts used to caution that Texas can never have too much water (true) and that any flood in Texas is at best brief because usually long droughts follow. The data tell the story and she's right!

(Her observation along with the oil crash provides an answer for why some people could get extremely cheap land in Texas during the 1980s.)

My summarized write-up of a longer research I wrote privately - https://sqlinsix.com/research-2015-texas-drought.

August. My Golden Asteroid story:

"In 2015, I gave a talk at a cryptosphere discussion on gold. My talk addressed how people around the world tend to view gold and why a cryptosphere project addressing this at the time (Digix) made sense.

One audience member remarked that gold wasn't rare because it was all over asteroids. You will find this belief common within the cryptoshere community. Since he made an assertion, I had a few questions for him about his assertion.

I asked him how much gold existed on the average asteroid. He didn't know.

I asked him how many metric tonnes of rock are needed to be extracted to get 1 troy ounce of gold on Earth. He didn't know.

I asked him how much would it cost to land robotic mining equipment on an asteroid. He didn't know.

I asked him how much would it cost to transport 1 metric tonne of gold from an asteroid to Earth. He didn't know.

I asked him how much the latest moon landing cost the United States - this felt like an easy question to answer for an American. He didn't know.

I would say to you what I said to him, 'You are the product of search engines and 24 hour news media. At least your grandfather would have been wise enough to not make the assertion in the first place.'"

November. Thomas Sowell notes in Applied Economics that nothing is more certain than risk. Is this how people tend to speak about risk?

December. Not the biggest fan of the company mixture with iShares RING, but at least gold miners have an ETF! Platinum mining has none.

2014

January. Todd - "You are at the center of the most important industry over the next few decades. Data will be the moat for the firms who get exceptional results or don't."

Todd can sell a vision. He made many good points about how firms that apply data principles well will excel ahead of their competition over time.

February. How to copy a table from one database to another database in SQL Server - https://www.youtube.com/watch?v=7ZMrw8lgYPo.

March. I hate pivot tables. Now I can just use my script in this video and viola! I have what I need - https://www.youtube.com/watch?v=ENC3oela9_E.

June. "We like the idea of creating code.. not managing it." Very common actually, especially with the ease of creating stored procedures. Here's how I find unused SQL Server stored procedures - https://www.mssqltips.com/sqlservertip/3243/finding-unused-sql-server-stored-procedures-with-powershell/.

In general, prevention is worth a pound of cure here. How much does it cost to remove something you don't use once you replace it? You can store and manage historic code, but keeping that code in production when you don't use it can cause confusion. Once you replace code or make it irrelevant, find better ways to retain it that don't interrupt current efforts.

July. Transaction (Process ID n) was deadlocked on lock resources with another process and has been chosen as the deadlock victim. Rerun the transaction. Brief video on this - https://www.youtube.com/watch?v=cGtZvEIVEDc.

August. Tech debt and the 20% philosophy - https://www.youtube.com/watch?v=-HFwfbx95JI. Easy to understand, but hard to do in practice. I observe that most people and firms will say, "Yes, we should invest some of our time in improvements, but.." It takes a commitment to follow through with investing 20% of time consistently.

September. Andrew Carnegie by David Nasaw. Rockefeller, Ford, now Carnegie.

December 31. Teaching my first technical course, Automating ETL, on Udemy. I hope this reduces some of the manual development approaches I constantly see. I'd also love to see developers completely cease using SSIS.

Earlier this year, Ricky showed me these SSIS packages that simply called stored procedures or executed T-SQL code. What is the point?

I also saw a lot of bad patterns when I did the latest round of interviews with candidates. I felt pretty shocked how many people couldn't get very far in our code challenge. I expected creation of a few dimensions or facts to get them, but it was far simpler. I think one reason why involves people not zooming out a little.

2013

January. How secure are private messages?

February. Second time I've needed to use this - https://www.sommarskog.se/query-plan-mysteries.html. This is a must favorite.

March. Even if financial companies can store data easily, don't forget that well-established firms avoid risk. "Bitcoin" to them feels like a ponzi scheme, not a price point worth storing.

April. Courtney - "He got an F because the teacher didn't like how he solved the problem, even though he got the right answer." I've heard similar stories multiple times.

May. Greg said something worth remembering - "I look back on code I used to write and hate it." That's how growth feels.

I don't study demographics the same way I did in 2009. I can also see all the assumptions I made that were wrong. I'll bet I also feel the same in a few years from now too.

On the code observation: thoughtful code tends to age well over time. Coding doesn't involve writing code. In my view, coding requires 95% thought and 5% writing it. When I see a block of code that integrates ten libraries, some of which cost, then I know that will be re-written at some point. When I see something like a state machine than handles every case without additional complexity, I know I can re-use that pattern with other problems.

June. I'm amazed how many people use SSIS to do the same thing BULK INSERT does because they're too afraid to write code, even though code is 10x faster and easier to maintain. You only have to write a function once, if you write it well.

A short how-to for myself since I use this all the time - https://www.youtube.com/watch?v=jd8bTGPJzAY.

August. Nicole - "He's pretty much stolen your identity." That's a phenomenal neg.

September. Sean - "[Financial leader in the United States] is my favorite communist." That's hilarious.

October. Earlier this year, I did a security webcast for SQL Server professionals. In the webcast, I zoomed out on security overall by advising firms to avoid storing data that they wouldn't want compromised. Cost would ultimately drive this decision.

I noticed how many educated people pushed back on this. Educated people like to think they're smart. They also like to think that they should be able to do things that they shouldn't. They remind me of the Soviet bureaucracy. They have a view of what they want to do regardless of what reality says. They impose their value even when reality shows they're wrong. They double-down, but make it worse in the process.

One commenter made me chuckle a bit because she provided a self-defeating argument. "If people store their own data, then an attacker will just hack them."

I rebutted that she missed the point. A hacker attacking and compromising 100 million people because a company stores all their data in a specific format doesn't take as much work as the attacker trying to attack 100 million individually. Not only are people extremely different, some people are also extremely disorganized.

The irony of disorganization in the context of security is that it can actually encumber an attacker because it doesn't make any sense. That gives the attacker pause - "This doesn't make any sense." No, it doesn't, but people don't always make sense.

When you zoom out, you see that the attacker ends up expending resources that reduces the overall attack space. This does not happen when an attacker compromises a company with 100 million users. Plus, the more users a company has, the more the incentive.

Security should fundamentally mirror how Mother Nature uses genetic variety and competition. Attackers are like diseases; they can wipe out a non-genetically diverse population easily. But a genetically diverse population? The survivors' genes become the new template.

The current security approach completely violates data in the feedback loop. Reality providing feedback means that you take the feedback, not deny it or pretend you could have taken another route. Fundamentally people will learn that they can't secure digital information. The digital world inherently lacks security. People make the choice for digital stuff because it's convenient and feels easy. But easy has never been secure ("easy come, easy go").

In my view, this also shows actual education. Education isn't a degree, certification or award. Education actually means you take the feedback reality provides and apply it. The correct way to secure priority data is to never store it in the first place. That is what reality continues to show.

2012

February. Robert sent me a link to the song Against the Wind by Bob Seger and noted how much he enjoyed the piano in the song. The song demonstrates a deep understanding of individual value.

March. Classroom sign: "100% of students will pass biology." American education summarized.

April. One of the drives failed, so I've been digging into drive capacity testing. I question whether any of these data would be useful to predict a drive failing in the future.

A database integrity check may pass, but not be restorable when on a questionable (or failing) drive. The same for both the backup and restore drive. If the backup drive is in a questionable state, it may pass the backup and fail restore on a good drive. Or it may fail the backup in the first place.

I can never be fully sure that any backup is good and that includes backups I've restored and run integrity checks. What I can know is that at least at one time the backup was good.

May. Josh - "You're very single-threaded." Noted.

July. I made a mistake today on a delete without a filter. I felt a little nervous about asking Nate for help, but he shared something that was really helpful - BEGIN/ROLLBACK/COMMIT TRAN. He also pointed out that because I write code, my mistake was easy for him to reverse. He hates the people who use the interface because they often don't know what they did wrong and he ends up searching for the mistake.

The big takeaways are to first query the set to be removed and use this query to get the data set, but also confirm after the removal. Second, use the transaction logic. Third confirm the records were removed. Fourth commit the transaction.

Nate also suggested saving the script for reference by date ("a script is a receipt of what happened").

August. Since when is Bambi a girl?

September. I'm surprised how many people speak about storing demographic information with personal and private details. After I completed my first major demographic study, I categorized groups, but avoided any identifiable information.

November. In my view, there's no science to indexing OLTP tables. The art is balancing write with read speed.

Final Note

One intended effect of public social media centers around both the feedback and encouragement of select content through the use of favorability and interaction. This affects how humans perceive content value. It further incentivizes participants to produce content to receive favorability and interaction over accuracy.

This page eliminates both of these incentives at the cost of perceived lower value by some readers. However, individuals actually determine value through its accuracy and utility for their situation. For this reason, any reader should only evaluate anything on this page from this lens.

All written content is copyright, all rights reserved. None of the content may be shared with any artificial intelligence application. Some content below this may appear later than when it was originally produced or may only appear for a period of time.

© 2004 - 2027 SqlinSix. All Rights Reserved. None of the written content may be shared with any artificial intelligence application.