... but alas, it was a full join. Thus, the glasses from those two tables duplicated and then duplicated again. And again. Soon, the glasses were filling the entire bar, from floor to ceiling. A union guy came and cleaned up, then declared "No more joins! If you want stuff from different tables, only unions are allowed to do that from now on."
Another great SQL history video: I have a 35 year career in SQL, and again, I was unaware of much of this history detail. And you managed to discuss Codd without going into his Rules, for which you are to be congratulated! I worked briefly at the Microsoft campus shortly after SQL Server 2000 was released, as as a customer, we were struggling to get an enterprise system ported from the earlier Sybase-based database engine code that still resided in SQL Server 6.5 into 2000, and performance was proving a big problem: the optimiser had been significantly re-written since Microsoft had forked the database engine following their split from Sybase. While I was there, I was told a story of what happened when Microsoft and Sybase parted ways. There was an agreement that upon termination of their cooperation agreement, all source code would be shared between them at that point. As you can imagine, Microsoft had done a lot of development separately. So when Microsoft sent Sybase all their final source code,, they stripped out all of the white space and mangled all the identifier names. I've never looked upon Microsoft in the same way since.
@@CaptainDangeax What exactly does the OS have to do with databases? I went from Linux to SunOS then Solaris, a stint with both Irix and AIX and on to MSSQL, ASP and .NET when it still didn't fully know what it wanted to be, stayed as a dev and systems designer with .NET and MSSQL with all the various subtrends of sorts until .NET 4.5 mostly. You can basically do anything you want faster, easier and with less resources on MS platforms compared to Linux. Linux is fine if you either opensource and do it all as a hobby more or less, or have an army of developers and testers along with a huge company doing something else too, like hardware. If you need enterprise level software developed quickly with constant iteration and customer responsiveness with a team of 2-3 people or just 1 - Linux isn't an option for a number of reasons but more than anything else speed and availability of outside resources to be hired in fast for specific tasks. But - each to their own I suppose. I'm not biased based on any other factor than what I am able to get off the ground into a full enterprise class product on my own reasonably quickly. Many times I've had that a new thing comes up, I sit down with a customer to understand their need and from that to a fully usable solution including a CRM, end user facing website, logistics partner integration, etc it's a matter of two weeks, maybe three - then it all goes live. Obviously after going live things come up but you deal with them as they come in, and 2-3 weeks after going live all issues are worked out. It then runs in production for years until there is a new something that needs to be built into it.
I remember an interview with Ted Codd in an 80's computer magazine where he made the interviewer sign an agreement that included the stipulation to refrain from calling him a "guru". Now I can't think of Codd or even SQL without that word "guru" popping into my head. Streisand effect.
we could have so much more great software if only we had those rabbits. makes you wonder whether sticking a CS-degree on everyone with a pulse was such great idea.
The last 25 years of my work life before retirement were spent in SQL. Really liked working with it, especially the ability for ad hoc queries when unexpected information is needed. Periodically management would need some very specific information, and SQL was a great tool for that task.
I have to say the amount of effort you put into the script and production is excellent. Having worked at Logica and used databases i really enjoyed this. These subjects are fascinating because although they are considered obscure they are literally world changing. Thank you.
16:00 Date's book is what my university used for the undergrad database class in the late 80's early 90's. I still have my copy. We used DB/2 on DOS-based PCs, and Date's book. This combination started to show the idiosyncratic differences between the different SQL dialects. Over her career, my wife has used Sybase, Ingress, SQL Server, Oracle, Informix, and DB/2.
Seventh edition dedicated to the 25-th anniversary of the first edition is right on my table now :) That's why I've moved to DBA after 20 years in IT - it is a most stable and conservative branch of the CS.
RIP Jim Gray. I had the pleasure of hearing him speak multiple times in my career at Microsoft before he was lost at sea. His disappearance sent shockwaves through the company. He was such a bright light in a company full of brilliant people.
@@fensoxx While Oracle was reasonably effective at its job as a database, its efficiency and stability didn't quite live up to the hype in sales presentations-it worked amazingly on the projector slide deck (the presentation materials), but customers using the tech in the real world didn't get the performance they were hoping for.
Incredibly well done history. I am developing a query system (not relational, but somewhat related) and I find this extremely informative and useful. Thank you!
This is like a trip down memory lane, im to young to have known any of these products from their initial stages but I have literally had the "pleasure" (millage may vary) to work with every one of the databases mentioned in this video. It's really interesting to hear how all these sometimes nightmarish products are related. Love the video's they are a must watch pretty much regardless of the topic being covered they are always very well researched and informative, truly a joy to watch. Though some like this one make me fee very old 🙂 (I am not all that old just yet, just had the "good fortune" of working for a large enterprise that had swallowed up a little over 200 of its smaller competitors within less than a decade. That is how I got to first consolidate all the different companies systems into a few central data centers to then start and consolidating the different business applications covering the same processes into a single system either existing or newly build. Being part of that work results in you working with pretty much every DB under the sun as the logistics industry which was where this organization made its money is not known for their progressive IT procurement policies.
I was one of the 500+ first porting engr and employees of Sybase in Emeryville, CA before going public. We definitely competed against Oracle. Worked right beside Bob Epstein. Greatest time of my 40yrs of software engineering!
Sybase used be in Emeryville before they moved to Dublin. I used to work at Oracle & Ed Oates office above my office & unfortunately I get his mail & he will get mine so I knew pretty well. He really didn’t need to work but wanted to do something.
Takes me back to my old days in Austin, Tx working at a software development company. At that time the company had made some pretty neat, small and fast software/browser-based emulators for the VAX, SYS 360, Unix/Linux and IBM 5250 Mini - all of them really including Amdahl and others from Europe. That was a lot of fun and I got to play with some of those mainframes and mini databases. They were remarkably fast and stable. It is amazing to me that SQL is still used to this day. A testament to its well thought out design. Who knows what the next innovation generation will bring? Perhaps something akin to using Quantum or other alternative systems with the only way I can put it, already connected so in a weird way "instant global connectivity and access to the data." Always enjoy your nostalgic content!
Hey your videos are really great, and I'm not even into SW stuff, I am an accountant that had to learn SQL when working in cost. But your videos are rich in history and I love history in all its forms.
@@vulpo Besides the CAP, noSQLs are popular mostly because the declarative approach of SQL is not very clear for coders stuck with imperative languages... Queries they write are terrible.
@@kondybas Yes, this is why Object-Relational Mapping tools like Hibernate (for Java) became popular [yuck!]. However some might prefer a simpler, lighter, and more transparent approach such as with Apache DbUtils that allows the programmer to have complete control and understanding of their data and SQL queries.
I remember reading long ago that Sybase would not sell to Microsoft, so Microsoft instead hired all of their top talent. An act that kicked off the popularity of non-poaching clauses between companies and non-competes in employment contracts. We're seeing that come full circle as more and more rulings and legislation (especially in CA) are dismantling these practices.
Hey bro, Awesome videos, I look forward to seeing them but I'm not sure what Google (RU-vid) is playing at but I just had to resubscribe to your channel.
as a DBA in SQL Server, that video is amazing, I also had the pleasure of briefly work with the IBM verson of Informix, in the Avaya CMS database, kinda different dialect of SQL, but pretty easy to use
Really like this story. Feels like I have been part. I started as DBA on mainframe network database - then was sysadmin on Unix with first version informix growing with two phase commit and SQL. I remember Stonebraker came to Informix . Anyway great memories - today we came so far with technology and this knowledge is as important as ever ❤
It would have been worth going into much closer detail about Oracle's early days. Just like IBM, the VC community believed software existed as a give-away used to sell hardware, and nearly everyone turned Ellision down when he tried to raise VC funding. When he finally succeeded, he got the brainless oaf Don Valentine on his Board, and Valentine was famously addicted to his notion "you can never fire a startup CEO too fast" - which led Sequoia to destroy dozens of potentially world-beating companies early in their lifecycle. Somehow Ellison managed to avoid Valentine's destructive impulses long enough to reach the IPO. Then, at the $1 billion revenue mark, Oracle nearly detonated because of Ellision's very poor financial management. Plus, Ellision played some very dubious games with the stock, resulting in his co-founders becoming nowhere near as incredibly rich as he himself did. It's an interesting tale and shows that there's a lot more to success than technology and timing.
It's amazing how IBM shoots one more time themselves on the foot. Another was Sybase selling the code to Microsoft, just to see Microsoft eat their lunch. My guess is they were thinking windows is never going to dominate the Enterprise, back then you had to restart Windows NT one time per day, not exactly stable. But they launch to the sky riding the WinTel model.
interesting how dedicated Oracle were (at the start anyway) to being compatible with Big Blue. That seems like a canny choice for an underdog. Double edged sword though if you accidentally become the industry leader.
You’ve missed a huge part of the history. That of Information Engineering, as promulgated and taught by Clive Finkelstein and James Martin. They developed a methodology for designing relational databases by modeling business data using the relational model. This was a prerequisite to implementing with Oracle or similar dbms. I know. I was there
I made my career doing SQL mostly, is a fascinating story. I wonder if Doctor Codd imagined how impactful was the whitepaper he wrote all those years ago.
Didn't Asianometry already do SQL a few weeks ago? For those SELECT few who read this, I say: AS FROM WHERE Asianometry got the idea to do another video on SQL, I'm not HAVING it. AND BETWEEN you and me, Asiometry, you can DELETE this post. But I LIKE and subscribe to your channel.
I read CJ Date's books when I learned SQL at UCLA. I used Microsoft SQL later at Massachusetts General Hospital to build a research registry for the neonatal ICU.
@@snuscaboose1942 No. HTML for the user interface, Cold Fusion 1.5 for the interface generator, and Microsoft SQL for the database itself. I could have used Javascript to make the menu mask prettier, but artistry is not one of my strengths.
20:59 Alex. Brown & Sons! Have to say I do miss the niche bankers like them, H&Q, even Adams, Harkness & Hill. At the time the financial side was more relational too (for better or worse). Seems less so now.
I have used every kind of databases in my career of computing. Every. The relational DB is by far the easiest to write software for, but also slowest of them all. Back in the day when the computers and their drives were veeeery slow, other types of databases were necessary. Today we can work with relational databases.
I never used SQL because the "relational" databases I was using was keyed with the realtime stamps and those were (practically) never in sync. Timestamps as double-precision real numbers, you have to interpolate something to match them, usually I interpolated the vessel position as we knew it travelled more or less in a straight line.
With the right hardware, RDBMSs out pace other systems in high concurrency ACID compliant transaction processing, such as core banking, share trading, gaming (horses), lotteries etc... They are also great for CRMs.
I remembre building an early version of Oracle on a Taiwanese AT-class compatible PC, back in the 80... It tooked some 20 floppy disks, at least ! But the product was running, complete. All on a 640KBytes RAM (I had to check !) machine. Waouh !
please do a course on Coursera or Patreon on anything you find interesting, especially on the actual history of technology, I would pay for it. alternatively, write a book.
@@louwrentius and for musk and zuck as well..and google, and bezos of course..all seed-funded by the usual suspects..and getting big fat contracts for DOD and the dozens of other services...
Reminder: The word "relational" in relational database has nothing to do with relationships. In mathematics a relation is a subset of a cross product of sets. The power of the relational database arises from its mathematical foundation. SQL, on the other hand, was just made up as they went along.
You gotta help me out with the English language here! What is a ship. In German both are called “Beziehung”. Is ship a state? Marriage or prof student relation are typical textbook examples . Like you would have one table with rows made of ( to keep things simple and typed ) wife and husband names. Or a table with many duplicates of the professor name. Note how in a community Names are made unique by adding a Roman number in case.
You do know that SQL is based on tuple relational calculus? Different vendors have variations on the standard, but a standard for SQL does exist and is maintained, ISO/IEC 9075.
I was confused by the image of the Sybase office. Didn't look anything like Dublin. Then I Googled it and realised it was Dublin in California, not Ireland! 😂