Beyond BI
-
One of the big news items from last week was Bill Gates memo on BI, or more specifically moving beyond BI. It's a similar talk that I've heard before about how we have information overload and underload. Too much information available, but too hard to find it.
While I agree with that part, I'm not that I think that BI or many of the technologies touted, collaboration, etc. will really solve most of the issues. People use email and search because it's easy, convenient, and they're involved in it everyday. IM use has grown because it is also very convenient and easy to fit into your workflow. However most other applications, despite being well designed to display information and perform a task, haven't grown quickly because they aren't intuitive or don't fit, or just plain aren't as easy to use.
I'm a tool guy. And more and more I find most other people are tool people. In other words, we increasingly view the computer as a tool, not a gee-whiz cool amazing piece of technology that we want to dig into. Even lots of IT people just want the computer to do the thing they need without messing with it. Email works like this and that's one of the great things about Outlook. You click a button, compose a message, click or type an address, and click a button to send.
Having to navigate another interface or even another program to collaborate doesn't make sense. I see where Sharepoint helps here, but the simple work with a document or message and then file/send/trash it model is what has been done with paper on desks for years. Trying to force us to evolve to some new model doesn't make sense and I'm not sure it will work.
Human interactions haven't changed a lot in thousands of years. We may be able to talk on a cell phone at the ball game or "communicate" with our Blackberrys, but it's still a conversation tool that's easy to use. That's why instant messaging works great. It's like a conversation between two people.
I'm not sure how much better the tools will be and Mr. Gates email isn't that specific about how they will improve the tools. I'm just not sure that BI is something we'll ever see implemented that widely.
Steve Jones
-
Let's just hope that people don't stop thinking as a result of all these upcoming innovations.
Just because we are shown an answer does not mean we understand it.
I have seen many an example of people touting graphs and tables in meetings and making it painfully apparent that they have no idea what they are talking about.
I just hope the large software developers remember this when developing their designs.
*plinks in his 2 cents*
-
Assuming they have developers. Saw yesterday they told ~1,000 contractors to take the week off without pay as cost savings.
Not encouraged at all with the overall performance of microsoft.
Greg H -
Ah, but remember this is a 10-year plan. I couldn't access the third page of the article, but Bill G is known for his ability to envision a future and then steer his company into that future. Love 'em or hate 'em, Microsoft controls a huge section of the software market, and will continue to do so.
If the planners and developers of this new software remember the basics of human nature, it is quite possible that it will be widely accepted. Remember, it's the 15 and 20 year olds of today that will be using it in 10 years.
That is, if they can deliver it on time!
Here there be dragons...,Steph Brown
-
I just went to TDWI data warehouse conference in Chicago. It was a wonderful conference talking about data warehouse and BI. At this point SQL Server BI is behind Oracle, Taradata, Business Object and some other tool. But with Microsoft ambitions, it will lead the BI world in no time. Actually there was one of my instructor's comment.
If you are a real data warehouse / BI developer, a tool cannot help you to build the data warehouse. You still have to collect information from the business, design the data warehouse, design the reports, manage the ETL process, and manage the performance of the data warehouse. The tool is just helping you to get things done easier.
Many people in the conference bought all the expensive tools under the sun, do they have the best data warehouse ? No. I mentioned to one of my instructor who was a leading data warehouse expert that I am using DTS for ETL tool. He said DTS is ETL tool and is fine if it worked for your company. Spending million of dollars does not help you to solve the problem and build the best data warehouse.
Just liked if you are a bad driver, buying a hummer or a BMW does not turn you into a good driver !
Just my two cents.
-
BI is already ubiquitous - think the grocery store "discount card" that is actually cataloging your purchase history, then printing related coupons on the back of your receipt for your next visit.
And take this with a grain of salt, but I view a tool as something that's predictable and reliable ... and a computer rarely meets that criteria. It barely meets a simple warranty of merchantability, to me, sometimes.
To use the old analogy, you would never stand for your car functioning with the same reliability as your computer.
And I am an IT professional with 22 years in the field with tech level knowledge on mainframe, mid-range, and micro computers. Some days it's such a comedy of errors, the power off button is their best feature.
-
Surely this is about context?
Anyone who has had the briefest exposure to AI knows that natural language parsing is not anywhere near good, sometimes it can be smart, but it isn't clever. As I see it there haven't been any major leaps forward in this since ELIZA in the 70's.
And there's the rub, as loner says, doesn't matter how good your tools are you've still got to make the leap from data to information by applying context to it, and this is a serious amount of fuzzy logic, at least as much as language parsing, maybe even more.
As the main editorial says, one approach is to get the users to change their habits, but this isn't going to work, never has, never will.
I'm sure a lot of people have tried simple things like getting the users to put tags on their documents for indexing and searching, and like me have found this just doesn't work. Not that anyone ever reads anyone else's work anyway, at least not unless they absolutely have to.
Well we're just getting around to solving these problems with things like google desktop and LAN search engines, and that's just the problem of understanding document content at a very shallow level, data context is many magnitudes more difficult.
No, sorry, don't see an answer to this until databases evolve another order of magnitude, probably akin to a leap from flat file to relational.
-
One part of the Gates editorial which struck me was the comment that we tend to over estimate how much will change in the next two years and under estimate the change over the next ten years. This has to do with the exponential impact of improvements in technology and technique. The first attempts to do most things are not great successses, but we learn a great deal from those attempts and go on to second attempts which are almost as bad as the first attempts. But with a series of slightly better attempts we eventually do the impossible. Does anyone remember how hard early attempts at WYSIWG document editors were? Now you have WYSIWYG in products as simple as WordPad. Voice recognizition software sucks now, but we have learned a great deal from our current attempts and continue to work at it. Parsing documents for semantic content is a daunting challenge, and maybe it is 10 or more years away from being practical, but evantually we will get there. And once somone solves the technical challenges, someone else will come along and figure out how to package it so that users can just turn the key and drive without understanding what goes on under the hood. Probably not going to happen in a year starting in 200, and may not happen in a year starting with 201 but it will happen and many of us will live to see it happen.
-
whoever wants to crack bi will have to successfully address two things: the behind the scenes, i.e. the reasoning engine, parsing or new technologies such as hawkin's htm (very interesting and exciting stuff - check out numenta.com), and the interface that users will use to interact with the system. the latter must receive as much attention, if not more, than the former. as the article mentions, the ease with which users interact with their tools is paramount to their adoption of new systems. google's personalization page is an accurate indication of where systems are headed. their build process is shorter, there is little to no installation required by the user and they're very intuitive. bi is the part of the system application that we will not see but will be widely use. i believe bi, or application intelligence in general, is the next growth area for data applications despite some recent consolidations. it only makes sense that, starting with businesses, users will want system apps to be predictive with the data they store and then respond appropriately based on those predictions.
-
That's why instant messaging works great. It's like a conversation between two people. I just can't bring myself to agree with that... I don't care how fast you type, it will never give you the speed of verbal communication and it will never properly set the tone of the conversation. Two people looking at the same thing and talking on the phone or side-by-each will accomplish in 10 minutes what will take an hour or two with all the text based, typing-required messaging you can handle. I think the best place for emails, crack-berries, pagers, and other non-verbal communications is from automated systems for trouble alerts so you can call someone to fix it. Yeah, yeah... you can distribute documents very easily to a large group of people via email but that's not what I'm talking about. If you need something done fast, have people talk to people at 600 words per minute instead of 20 to 60 words per minute by typing. Answer the phone, folks!
So far as BI goes... I've seen some better than others and, I imagine, someone will eventually get it right, but for now, my opinion is that Business Intelligence is an EOM (Expensive OxyMoron). Still, it's great for Developers... their managers have something to blame instead of them
--Jeff Moden
RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
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.Change is inevitable... Change for the better is not.
Helpful Links:
How to post code problems
How to Post Performance Problems
Create a Tally Function (fnTally) -
Jeff,
I think you're right with IM for those unfamiliar with each other is hard. But for those that work together and know each other, you'd be surprised how much communication comes through. For intra-team communication, it's a great tool.
'Course the conversations do need to be short. Can't type 2 paras and send them.
-
Am I missing something here? In Bill Gates Message, he was not just talking about communicating between co-workers. He was talking about the communicating within the company. Nowaday data is the biggest asset of the company. The senior managements need data to drive the business. What Mr. Gates talked about was BI (hopefully) would be the tool to gather the data within the company to help the senior managements. For example, if the CFO wanted to find out the sales of the month, he needed sales dept, finance dept and IT dept to compile the data. BI may be able to help to get the job done much faster and easier.
I am sorry if I missed the point of what Mr. Gates was talking about. Or I am sorry if I missed the point of what you guys are talking about. I am confused here.
-
Sorry... wasn't talking about the concept of BI... I was talking about the expensive products people buy to do BI.... we made our own because nobody seemed to be doing it right without charging big $$$.
--Jeff Moden
RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
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.Change is inevitable... Change for the better is not.
Helpful Links:
How to post code problems
How to Post Performance Problems
Create a Tally Function (fnTally)
Viewing 13 posts - 1 through 12 (of 12 total)
You must be logged in to reply to this topic. Login to reply