I've resigned myself to the fact that even the world's MVPs don't have much of an impact on what MS does. For example, there's been an "approved" CONNECT item that was approved almost 7 years ago now for a function to very quickly generate either a numeric sequence or a date sequence. Nothing yet in the product, though.
They also take steps backwards. For example, Pinned Tables where extremely useful for things like Tally Tables and other helper tables. They took that ability away saying that the system knows when to cache something and keep it cached. Here we are coming up on 2014 and the introduction of Hekaton... which isn't much more than Pinned Tables on steroids except it has more restrictions than the old Pinned Tables.
And they still don't have anything to speak of for file handling or data exports from T-SQL. The overwhelming suggestion is to use SSIS for that. Yeah... another system to maintain, another language to memorize, and more troubles than you can shake a stick at because it doesn't do it all. If it did, you wouldn't have to write bloody scripts to make up for what it can't actually do.
Then there's MS's idea of what should be deprecated and dropped or not. I won't even go there because it's such a huge subject.
What I'd like to see is for MS to get back to some grass roots programming to make T-SQL what it should/could be. For example, fix Bulk Insert and BCP so you really can skip a header that doesn't have the same kind/number of delimiters as the data. Make a Bulk Export that's as easy to use as Bulk Insert. Stop farting around with add ons like needing ACE drivers to import a freaking TRUE CSV file or a spreadsheet file. And fix PIVOT so that it's at least as it is in Access. The current version is a joke. Build that stuff in and make it easy to use and do. People have to do this stuff a whole lot.
The reorg should have been to split things up a bit... one to build new stuff (like Hekaton) and build it with some decent technology and no missing parts. Another group should be formed to fix all the crap (sorry, could help but call it crap) that's out in the field on a retrofittable basis. Lot's of people invest in MS technology like SQL Server and they keep the same rev for 6 to 10 years simply because MS makes revs have little "gotchas" from rev to rev that require full testing and lot's of rework because of deprecated features (decmial indexing, anyone?). How about the fact that you need to go to "Single User" mode for an entire database to set even just one filegroup for an Audit table to Read Only because it's static data!?
And holy shades of "Peter Norton", Batman!!!! Would it REALLY have been that bloody difficult to apply some 1980's technology when shrinking a database or rebuilding a clustered index???
Then there's the Office products... if they had spent as much time on giving people the choice of classic menus or using that (for example) bloody Ribbon Bar (which has really slowed me down a lot... takes 7 or 8 clicks to do what used to take 1 or 2 to get to the very same menus) as they did on "cool looking eye-candy" such as making "translucent" windows (REALLY??? How does THAT make you more productive???), they'd have a much reputation.
This reorg reminds of what happened in the 70's when RITZ crackers changed their box. They advertised that "Not to worry... You can change the box but it's still the same ol' Ritz!" and I'm really afraid that MS is going to follow in that path except their crackers can be pretty rotten at times and they need to change the cracker instead of the box.
is pronounced ree-bar and is a Modenism for R
First step towards the paradigm shift of writing Set Based code: Stop thinking about what you want to do to a row... think, instead, of what you want to do to a column.
If you think its expensive to hire a professional to do the job, wait until you hire an amateur. -- Red Adair
How to post code problemsHow to post performance problemsForum FAQs