|
Ah OK, my mistake. Thought you were talking about SQL itself, not the database.
I suspect 99% of databases live as long as the product they serve without significant rewrites.
|
|
|
|
|
pt1401 wrote: I suspect 99% of databases live as long as the product they serve without significant rewrites.
You are probably right. So I should add - the product has been updated 6 times (only major counted) since then... The original version was a desktop application in C and now it is a full web based including support for handheld devices...
Skipper: We'll fix it.
Alex: Fix it? How you gonna fix this?
Skipper: Grit, spit and a whole lotta duct tape.
|
|
|
|
|
pt1401 wrote: Windows was launched 35 years ago - has it been updated? Or upgraded? Downgraded.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
Nope.
Database-theory has not changed. You state it was 'designed 25 years ago'; the word designed implies that it has been normalized.
Perhaps you prefer entities without any upfront design? (As in, evolving by Agile)
Bastard Programmer from Hell
If you can't read my code, try converting it here[^]
|
|
|
|
|
Eddy Vluggen wrote: Database-theory has not changed.
But we got better tools to get close to it...
Like using SF, Func, Trigger, XML data (as XML data and not string manipulated by SQL string methods), CTE, new kwywords, like OUTPUT, built in paging and so... We got a lot of it in the last 25 years...
Skipper: We'll fix it.
Alex: Fix it? How you gonna fix this?
Skipper: Grit, spit and a whole lotta duct tape.
|
|
|
|
|
Triggers aren't new in the database-world, and XML is not exactly the ideal format for an RDBMS.
Kornfeld Eliyahu Peter wrote: We got a lot of it in the last 25 years... Yes. Doesn't mean it has to be used.
Bastard Programmer from Hell
If you can't read my code, try converting it here[^]
|
|
|
|
|
Eddy Vluggen wrote: XML is not exactly the ideal format for an RDBMS
Exactly. But it has been used - so please take a moment and move from string manipulation to faster and better options, now you have them!!!
Skipper: We'll fix it.
Alex: Fix it? How you gonna fix this?
Skipper: Grit, spit and a whole lotta duct tape.
|
|
|
|
|
Kornfeld Eliyahu Peter wrote: so please take a moment and move from string manipulation to faster and better options, now you have them!!! Ehr.. databases existed before we had XML. So, the faster and better option would be - the database.
..and there will always be this bright genius who is going to store 32 flags as a bit-string, manipulating the thing with .Left$ en .Right$ until he gets his flag.
Bastard Programmer from Hell
If you can't read my code, try converting it here[^]
|
|
|
|
|
Eddy Vluggen wrote: this bright genius And I got this DB from that 'bright genius'
Skipper: We'll fix it.
Alex: Fix it? How you gonna fix this?
Skipper: Grit, spit and a whole lotta duct tape.
|
|
|
|
|
Tools can't fix that.
So, have you considered throwing it all away?
Bastard Programmer from Hell
If you can't read my code, try converting it here[^]
|
|
|
|
|
Eddy Vluggen wrote: ..and there will always be this bright genius who is going to store 32 flags as a bit-string, manipulating the thing with .Left$ en .Right$ until he gets his flag.
Most likely 16 bits (or 15 coz top bit made number -ve and messy).
Also smaller numbers applied to hard disks - measured in [often single or if lucky double digit] megabytes
memory was in kilobytes - everything got squeezed.
25 years ago all numbers were a lot smaller than they are today.
(But the best part: no damn cell phones, once ya left the office work was really over.)
Sin tack
the any key okay
|
|
|
|
|
Please stop whining:
Twenty five years ago they did know how to design a database.
|
|
|
|
|
Well, yes, 25 years ago there were some people who knew how to design databases, but there were an awful lot who didn't. Many people came to SQLServer or Oracle directly from ISAM systems, so it was hardly surprising that many didn't really grasp RDBMS.
Have things changed since then?
Maybe - I think that people are generally a bit more clued up than they used to be on this front but I can guarantee that even as I type this, somebody, somewhere is creating a "database" with mile-wide tables and not a single foreign key to be seen - and they don't have the excuse of it being new technology.
People who don't know RDBMS 1.01 (the fact that the 'R' stands for "Relational") are still banging together monstrosities that will be causing pain for the poor souls who have to clean up after them for many, many years to come.
Until the world comes to its senses and introduces the death penalty for denormalisation, this problem will continue.
Slogans aren't solutions.
|
|
|
|
|
PeejayAdams wrote: Until the world comes to its senses and introduces the death penalty for denormalisation, this problem will continue. Denormalisation isn't necessarily a bad thing, unless you mix it up with randomly-dump-any-value-where-you-feel-like-it "methodologies".
If you need a lot of derived values, for example, it can often be less costly to have them handled within the database, rather than build them into a front end, which may radically change, year over year.
Denormalisation can also provide significant performance boosts, especially where joins begin to resemble spiderwebs built by drunken spiders.
For me, the sole function of a db is to provide needed data as efficiently as possible. Most of the time, that means it's best if it's at least third normal, but "most of the time" isn't "always".
You shouldn't close doors to effective pathways.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
Well, yes, I agree that we need to be pragmatic and I was maybe being slightly flippant in suggesting that all forms of denormalisation, rather than most, should lead to the slow and painful death of the perpetrator
Mark_Wallace wrote: If you need a lot of derived values, for example, it can often be less costly to have them handled within the database, rather than build them into a front end, which may radically change, year over year.
I'm not saying that I would never add redundant data but I'd need a rock solid case for doing so. In my experience, redundancy pretty well always ends in tears. Maybe not today, maybe not tomorrow but somewhere down the line ...
Mark_Wallace wrote: Denormalisation can also provide significant performance boosts, especially where joins begin to resemble spiderwebs built by drunken spiders.
It's very rare for this to be so; the optimiser thrives on foreign keys - it is, in itself, optimised for normalised databases. If joins are becoming spidery and problematic, I'd be looking for the cause of that before I considered reducing their number by denormalising the DB.
Mark_Wallace wrote: For me, the sole function of a db is to provide needed data as efficiently as possible. Most of the time, that means it's best if it's at least third normal, but "most of the time" isn't "always".You shouldn't close doors to effective pathways.
I completely agree with that! My general rule of thumb is 3/4NF in most cases, 5NF if the thing has to really fly but that is, very much, a rule of thumb and rules should never be immovable.
Slogans aren't solutions.
|
|
|
|
|
Tables is still tables, Shirley?
You might want to recook a few stored procedures, scripts, and queries, but unless the data has changed hugely and new table structures haven't been implemented well, it shouldn't be too painful.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
So only INSERTs and DELETEs?
|
|
|
|
|
Kornfeld Eliyahu Peter wrote: SQL 6.5
IIRC, wasn't that the last version where every database had to have its own dedicated partition?
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
I know what 365 means in Office 365...
You have no mail for 365 days in a year...
I can live with that
Skipper: We'll fix it.
Alex: Fix it? How you gonna fix this?
Skipper: Grit, spit and a whole lotta duct tape.
|
|
|
|
|
Does that mean that you only check your mail on Feb. 29?
If you have an important point to make, don't try to be subtle or clever. Use a pile driver. Hit the point once. Then come back and hit it again. Then hit it a third time - a tremendous whack.
--Winston Churchill
|
|
|
|
|
Office 365 currently has 'A problem with my account' that they 'need my help to resolve'.
Apparently I need to uninstall Office, install Office 32-bit, uninstall Office 32-bit, install Office 64-bit before it'll stop implying that I don't have a license.
I really can't be a*sed until it refuses to let me use it.
|
|
|
|
|
Also check your Outlook version - 2003 has no support without hotfix...
It is something about SSL/TSL support of Outlook...
Skipper: We'll fix it.
Alex: Fix it? How you gonna fix this?
Skipper: Grit, spit and a whole lotta duct tape.
|
|
|
|
|
|
Hell, it serves you right for picking the worst ever version of ms office.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|