March 20, 2009 at 5:31 am
Steve Jones - Editor (3/19/2009)
...that barely needed any of the functionality of SQL Server. So many of them take advantage of bare features, essentially using a dumb data store.
We use an established investment accounting system. It was originally developed using Powerflex and old VB code. It sort of morphed over the years. Then the vendor announced a huge milestone. Their system now uses SQL Server!
And there wasn't much rejoicing. Their idea of using SQL Server is nothing more than a glorified OS. There are NO views, Stored Procs, etc. All they did was create tables that mimic the flat files of Powerflex, and attach an ODBC interface from their original Powerflex apps to SQL.
Things like pricing purges were taking 6+ hours to run. Our resident expert asked if I could do the same sort of functionality in native SQL. I took the specs and wrote a delete statement. Finishes in about 15 seconds.
How about this "Feature", their system can not identify records that have changed. That's just the tip...
Cant wait for them to try the next version of SQL... One of their favorite things is going away. Direct R/W access to system tables.
Now the scary part. There are only about 3 or 4 packages available, and this is the best we can get. (The best in the industry isn't available to us, because it was/is developed by a competitor.)
Honor Super Omnia-
Jason Miller
March 20, 2009 at 5:43 am
Wow, that is scary! And programmers get paid to write this pile of %#*@
You could always suggest that they save money by migrating to mySQL, or PostGreSQL, or even back to flat files 🙂
March 20, 2009 at 6:01 am
AndyD (3/20/2009)
Wow, that is scary! And programmers get paid to write this pile of %#*@You could always suggest that they save money by migrating to mySQL, or PostGreSQL, or even back to flat files 🙂
Cant. The vendor dictates the platform. It took an act of dog, or more specifically, our head legal counsel, to get them to agree that I could create triggers on their tables. Now I can at least track what has changed.
Not only do they get paid, they're paid VERY handsomely. (Well, the company is, I don't know about the developers.)
We're talking in the millions of dollars here...
Honor Super Omnia-
Jason Miller
October 29, 2013 at 6:12 am
Interesting reading this topic 4 years on now.
I found Steve's comments that most of his instances pretty much used the same features of SqlServer interesting. In the same vein, we've got a server which runs SS2005 and while there are some features of 2012 that would be useful, we're not at all convinced it's worth the expense, considering the features we use and MS's head-spinning pricing structure.
The funny thing is I discovered another instance of SS2005 on a different server that's been sitting unused for quite some time - it's still at the RTM version (!). Needless to say that one got shut down until we get a chance to bring it up to speed and/or decide what to so with it.
Now at home I have 2012 Express for tinkering around with and will probably spend the $60 or so to get 2012 Developer.
____________
Just my $0.02 from over here in the cheap seats of the peanut gallery - please adjust for inflation and/or your local currency.
October 29, 2013 at 6:32 am
What's different today? There's a lot of people using smartphones and tablets that aren't Windows based. While my workstation and most of the servers at work are Microsoft Windows, we have a significant percentage of Apple and Linux machines.
My personal laptop and home workstation run Linux and run it well. I keep a Windows virtual machine on standby for the periodic apps like tax prep and work of that ilk. There's a whole set of OSS tools designed to handle data that are worth looking into. It's an option for those that aren't fully committed to a single vendor.
October 29, 2013 at 7:22 am
This article was published back in 2009, and four years down the road, and SQL Server is still going strong. The arguments sound similar to what some folks were saying about Microsoft and web-development back in 1999, and then .NET came along. It also sounds a lot like arguments people were saying back in 1989 about how Microsoft didn't get GUI operating systems, and then Windows materialized leaving OS/2 and Mac to flounder in it's wake for years. Microsoft is actually a master when it comes to knocking off the competition. When the industry says that Microsoft doesn't get something, then get ready for Microsoft to prove them wrong.
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
October 29, 2013 at 8:34 am
Microsoft is actually a master when it comes to knocking off the competition. When the industry says that Microsoft doesn't get something, then get ready for Microsoft to prove them wrong.
How are they doing in the smart phone/tablet business? Or being overpriced on their new gaming console? How's the corporate acceptance of Windows 8.x? Has the number of supercomputers using Windows increased? Or the number of embedded systems? Or the number of web servers? How about the price increases on MSDN or SQL Server? The dropping of TechNet and XNA?
There's a few key pieces that make Microsoft essential to many companies and people. Office, Active Directory, the Windows infrastructure, SQL Server and Visual Studio. Ask yourself which of these can be replaced or scaled back by an individual or company.
SQL Server is overkill and expensive for smaller organizations. As much as I like it, many of my clients don't need it, they can be served by either SaaS or by an OSS data solution. Like I said, there's more options now than four years ago.
October 29, 2013 at 10:22 am
While there may be some SQL Server knockoff someday doesn't mean it will be the end of SQL Server from Microsoft. Case in point, there has been a system for a few years now called EnterpriseDB based on PostreSQL that is Oracle PL/SQL compatible, yet acording to this site's research is only ranked 37th among relational databases in popularity. (even if you don't agree with their methodology in calculating scores, it's still interesting)
http://db-engines.com/en/ranking/relational+dbms
Just because something is becoming or already is a commodity doesn't mean that there isn't a market for specialty or high-end versions of those products.
October 29, 2013 at 12:02 pm
"One of their main products, Office, is essentially done. What more can you do to those applications that people need?"
I liked this sentence from 2009. I know for most regular people Office today is the same as Office 95, but I think if you really had to go back and use a version that is 2 behind the one you're used to, you would notice it.
I'm using Office 2013 now. If I had to go back to Office 2007 I would miss Power Pivot, Power View, Quick Explore. A ton of little optimizations that I've gotten used to and I don't think much of, but haven't always existed. If I had to go back to Office 2003, I'd miss OneNote. The docx file format. Custom sorting in Excel. I'd even miss the much-hated ribbon interface.
Over time, "new" features you didn't need become standard.
Leonard
Madison, WI
October 31, 2013 at 2:34 am
Coming back to this after 4 years the observations I would make is that the world has taken a different path.
People needed a low cost alternative to Oracle, SQL Server, DB2, Sybase and to fill this gap PostGres and MySQL were born.
The fundamentals of PostGres (one of Stonebrakers babies) have found there way into a number of other mainstream technologies.
Microsoft have actually been behind the curve with column technologies as Infobright, as a column engine compatible with the MySQL query engine) has been around for donkey's years. For data in the dimensional model Infobright will significantly out-perform non-column store SQL Server and is writable.
Vertica (another Stonebraker child) was an open-source MPP column store. It works well with SSIS & SSRS. My expertise isn't in SSAS but I strongly suspect it would be fantastic for ROLAP. Community edition is "limited" to 3 nodes and 1TB of raw data. It isn't as fast as SQL2012 column store but its scale-out nature means that it can grow beyond the boundaries of SQL2012.
Then you get into the NoSQL camp. The big 6 (MongoDB, Redis, Cassandra, RIAK, Neo4j, Hadoop) are very different from traditional RDBMS. Although the market penetration is far less than would be suggested by the number of articles on the web they can fulfill many of the use cases that SQL Server and RDBMS's per se have had as defacto data stores.
Stonebraker steps forward again with VoltDB and Jim Starkey's NuoDB (formerly NimbusDB) look extremely promising.
Of course Microsoft have taken a big step forward with SQL2014 and the Hekaton engine. If you have a large and multifaceted data estate then SQL server is probably a good fit for you. If your needs are much simpler then it may be overkill for what you are trying to do.
November 26, 2013 at 2:23 pm
I have recently learned that our organization is getting an IBM Netezza data warehousing appliance for use as a replacement for one of our large SQL Server reporting servers. Executive management had briefly considered SQL Server PDW edition.
David, Steve, or anyone else have any experience or thoughts about Netezza?
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
November 26, 2013 at 4:13 pm
David.Poole (10/31/2013)
People needed a low cost alternative to Oracle, SQL Server, DB2, Sybase and to fill this gap PostGres and MySQL were born.
Presumably you are talking about a time after Sybase stpopped calling its product SQL Server and the name was used only by Microsoft, since you list SQL Server and Sybase as two separate things to which it would be an alternative. Even if you are talking about an earlier date than that, you have your history very wrong. The Postgres project was set up two years before Sybase released it's first odd bits of prototype, 3 years before first release of Sybase, and its first release was one 49 months before the first port of sybase 3.0 to Os/2 that MS had permission to do its own marketing of (as opposed to joint marketing with the other two owners of the product), that is about 6 years before MS first released a version that was more - only very slightly more - than a straight port of Sybase SQL Server. Sybase didn't change its product name from Sybase SQL Server to Sybase Adaptive Server Enterprise until 1997.
It's quite clear that Postgres wasn't a response to market pressure (Stonebraler had just released his commercial version of Ingres) but what Stonebreker said it was: a successor for Ingres in the academic world (in the commercial as opposed to academic world, Ingres was reasonably successful and still going strong last time I looked at it, in the early/mid 90s, but it's dropped back quite a bit since then) for people who wanted improved concurrency for OLTP and a decent user-defined type system and suport for Stonebrakers view of what domains are in the relational model.
For MySQL, the dates are a less impossible fit for your statement as MySQL 9 or 10 years later than Postgres. But when MySQL began both Ingres and Postgres were available as low cost alternatives to the big boys, so I doubt that the Swedes saw it that way rather than as a way to have open source rdbms project that conformed to their concepts or openness rather than to Californian concepts, or maybe just wanted something that was to a great extent mSQL compatible.
edit: while on historical inaccuracy, I can't resist quoting this from Steve's article:
What if we get to the point where a MySQL or PostgreSQL type knock-off that runs T-SQL comes along and implements 80% of what SQL Server 2005 has
Intersting idea: Postgres had snapshot Isolation before Microsoft SQL Server was first released, and had a very effective equivalent for what MS call CLR (much better, much easier to use, much more flexible right back then than IBM DB2's awful support of various languages for writing bits to be called from SQL was in 2000; it's hard to compare it as a whole thing with CLR because it's not based on a comon language runtime for a whole network application architecture in teh way that SQL CLR integration is, but in terms of allowing code to be built efficiently in a language other than SQL and then imported into and used in the database it is directly comparable). So things seem a bit more like SQL Server is a knock-off of Postgres that implements maybe 80% of its features....
Tom
Viewing 12 posts - 31 through 41 (of 41 total)
You must be logged in to reply to this topic. Login to reply