Slashdot Mirror
What Gartner Is Telling Your Boss
Littlewink writes, "Esther Schindler's latest analysis reveals what Gartner is telling your boss at their annual conference. Excerpts: '"The future of application development is not about programmer productivity," said [Gartner analyst] Hoyle during the keynote presentation, "but in assembling functionality from components." [Gartner analyst] Veccio stated "Why would you ever code an app from scratch again? Why would you need to?"' According to Schindler (who does not 'drink the Kool-Aid'), Gartner urges managers to consider better process control and governance, managing 'application portfolios' much as they do stock portfolios. Part of this discipline is 'killing development projects early and often.'"
Basically they're just another rent-a-quote firm for people who buy their services
Why would you ever assemble wisdom and business savvy when it's simpler and easier to assemble random quotes & concepts from popular seminars and "Best Seller" managerial books.
No, really. I worry about the junk that comes out of Gartner. Like the outsourcing binge. Where I work, they have tried several ways to outsource (overseas) our work. Gartner was flogging this heavily and it seemed to become 'cause celeb'. The word we got from our managers was that it wasn't going to be allowed to fail. I'm pretty sure some people's careers would be damaged if it did, so they're going to continue to push it regardless of the results. This sort of 'tell us what to think' mentality is not going to help corporate amerika.
Please, someone tell Gartner Group to be a little less certain about their predictions. The mass of middle managers are afraid to do think anything that isn't supported by someone like Gartner.
Best regards.
Of course, the savvy OSS officiando will explain to management that this is exactly what they are striving for: "WE only write new code where strictly necessary ... most of this project is utilizing reusable components generated by the open source community..." etc.
It's an outsourcing thing. You can usually get away with offshoring glue code, but not major component work.
Gartner is just trying to justify offshoring and make $$ by telling MBAs what they already believe.
The best solutions to specific problems are going to be custom made, at least for a while.
Yeah, but sometimes you have to gird yourself for those days when the sheep come home from being fleeced at the latest management fad sheering.
I vividly remember the epic battles that took place when managers returned from TQM (Total Quality Management) training. The all had these purposeful looks of the new acolyte and a Franklin Planner under their arm. They cooked up Vision and Mission statements and tried to get everyone on the bandwagon. It was a trying time because most of the way we already did things were obviously the most efficient. Work under the gun a lot and you tend to find the shortcuts yourself. If anything we became less efficient until the whole clamour died away and most of us returned to getting it done the proper way.
A feeling of having made the same mistake before: Deja Foobar
Remember.. your CEO is just business object--- I set of components and business logic. His job responsibilities are just more components.
Any of these business objects can be swapped out and replaced willy-nilly as you see fit. If the CEO has too much work on his hands, you can simply run a process scanner against his position-- the process scanner will highlight the areas for improvement. Then you hire a new person and assign some of the objects to him.
Heck? Want to replace the boss? Fire him and hire a new object to assume the responsibilities. The transition is seemless.
Don't forget that you paid some consultants $1 million for this study, and these are the conclusions.
---
Look-- looking at things as components is a useful exercise for modelling. It's an easier way to get a "big picture" perspective without getting mirred in the details.
But it will only get you so far, because DEVIL IS IN THE DETAILS. Anbody who believes in such object-oriented drivel is certain to go out of business. Trouble is, the CEOs who promote this crap can jump from ship to ship-- not all of us can do that.
Asking "Why would you ever code an app from scratch again? Why would you need to?" is like asking "Why would you ever want to have a baby".
Sometimes it's the only way to develop what you need; sometimes it just happens by accident; and sometimes someone gives you one to look after for them.
You don't want to have a baby very often, but it's just as well that some people have them sometimes.
We're thinking about throwing Java out. It has the same problems with 'synchronisation' that C has with 'memory allocation'. You can't get it right all the time, it's too hard.
And Intel are coming up with these 80-Core chips.
A real lot of stuff will have to be rebuilt if we do. Hopefully automatically built from modelling tools. But there will have to be people, to resolve the defects, if it is to support the business.
I sometimes get the idea that data modeling is one of least used methods for building information systems. I wonder why.
I absolutely agree. Data modeling is one of the most fundamental skills out there, but time and again I encounter apps with just an absolutely atrocious data model. Much more time needs to be devoted in school to the fundamentals of data modeling and the why behind data modeling best practices. Think about it -- in a "classic" MVC stack, the controller and GUI are often interchangeable, but if you're stuck with a poor way to persist data, the rest of the app *will* be quite limited no matter what you're using for business logic and / or presentation. Furthermore, none of these "component" vendors will help you . . . you'll just end up with a turd wrapped in Company X's duct tape.
"Why would you ever code an app from scratch again? Why would you need to?"
.net, no java), politics and empire building, and any number of other variables all come together to make the business environment so complicated that developers will be reinventing the wheel for years to come. Things like MQSeries or Oracle have gone a long way toward standardizing things. You don't cook up databases in flat files anymore. You don't (usually) write your own messaging systems by opening sockets directly. Things like the .net framework or j2ee mean we're not writing sorting algorithms anymore. But that just frees us up to work on other complex systems. And complexity is growing faster than these sorts of standardized frameworks can be created. We'll continue to use standardized middleware packages and other third party controls or libraries. But the business will always need custom solutions that build on those standards.
Assuming they mean business logic and not things like sorting algorithms, you had better have vast quantities of foresight to make that happen. As most other crud conjured up by these people, it sounds great on paper and when given to executive types in the form of powerpoint presentations, but in practice, it falls apart. Different programmers, different programming styles, changing business rules, mergers, new client requirements, scope creep, abandoned products, legacy code you're unable to get away from, new business standards (like we're a java company, no
Disconnect your television. Do your own research. Draw your own conclusions. They're probably lying. Don't be a sheep.
"We need an inventory module for this project. Use the one from the MegaMess project."
"Yeah, sure, that should work. A little overkill, but it'll do what we need."
"Oh and it has to track the inventory by supplier's parent company and supplier's parent company's SKU."
"Uh....what?"
"Yeah, we get most of our inventory from local suppliers, but they all get it from the OEM. We need to use the OEM part numbers, with an indication of which OEM it is."
"Uh, but the MegaMess project tracked inventory by product group. It doesn't even use SKUs. And we don't need to report on SKUs for what we're doing, why do we need them?"
"Director of marketing wants to see a report broken down by SKU, and rolled up by parent supplier."
"I don't think we're going to be able to use the MegaMess inventory."
"DAMMIT! Just use the components we've already got! We aren't going to write any new code for this!"
Just junk food for thought...
The 80/20 rule messes up the reality of using components (whether it's EJB/SOA/latest cool thing). It takes 20% of the time to do the easy 80% of the work. Then 80% of the time to do the remaining hard 20%. Components give you the easy 80%. Which you could already do pretty quickly anyway, so you're really not gaining much.
Then you're still left with the remaining hard work, which probably got harder and will take longer due to the overhead of your component framework and its mess of configuration.
And that is totally ignoring the fact that it's very hard to find components to reuse anyway.
Nerd: Derogatory term typically directed at anybody with a lower Slashdot ID than you.
The layer of people between the decision makers and the product makers exist in every industry, and in every industry there are far too many of them, so they must justify their jobs and salaries. Isn't that more important than innovation or efficency? To them it is.
We are all just people.
Question: "Why would you ever code an app from scratch again?"
Answer: In order to avoid bloat, stability, performance, and security issues from using modules that are overly-used, overly-general and/or don't exactly meet your spec but are "close enough."
Best Example: Microsoft Office
What are the proposed alternatives that simplify synchronization?
(1) Software Transactional Memory, STMs. You write "int STM x=5;" and then one thread can do "atomic {x+=1;}" and another does "atomic {...;x-=1;...}" and the runtime+compiler magically make the atomic blocks execute atomically. This is different from Java synchronization blocks because these can be executed optimistically in parallel and because they never result in deadlock. People therefore say that STMs are "compositional" in a way that locks are not. The key is that the compiler/runtime know how to roll back an atomic block. This kind of atomic block interfaces very gracefully with SQL transactions.
(2) Message-passing. Academically it's embodied in the "pi calculus". In web terms it's embodied in the W3C "choreography" working draft. In practical terms it's embodied in Microsoft's Biztalk and in Ericcson's Erlang language and in Microsoft's new Robotics SDK. It's also a little reminiscent of "tuple-space" operating systems. The idea is that threads communicate by sending messages to each other. It's still possible to deadlock (e.g. if one thread waits for a message that will never come) but these errors seem more rare in practice. Also it's easier to analyse for this kind of problem at compile-time than it is with synchronization.
(3) Write the code in a functional way, so the compiler infers parallelism automatically and you don't need to. Or, take existing C code and have the compiler infer dataflow in it, then proceed as above. e.g. WaveScaler at UW, I think. Many people think this is the best hope for the coming world of multi-core chips.
In some ways I agree, 'Not Invented Here' is a real problem, where a department wants to come up with it's own tools, instead of using a shared set of tools.
There are some things you shouldn't do yourself, and some things you should.
Janitorial staffing, for example. Should each department in a building owned by the same company hire a seperate company to clean their offices? Obviously not. But should they all be required to use the same text editor, no matter if they are laying out advertisments or writing C++? I think not.
Both of those cases are obvious, but what about the text editor used by programmers in different departments? Unfortunetly, usually the person in a position to make company-wide policy does not know enough about a specific job area to make reasonable blanket requirements, requiring all developers to use a particular editor, no matter if they are developing for Windows or Linux would be like telling all janitors to use the same floor cleaner for office carpets and the parking garage.
In the Janitors case, since they are often outsourced, or at least a seperate department, they have their own structure which tells them what to use where.
In the software developers case, having a seperate structure to set standards can lead to problems when the Project manager's directions conflict with the standard practices; the project manager's desires usually take hold, because they are in direct contact with the developer, while the company standards are less strictly enforced. This leads to the effective death of the 'standards'. After this happens a number of times, everyone loses faith in anything labeled a company standard, and since they expect no support, they don't even really try to adhere to them.
I don't have a solution, as once an organization reaches this level of NIH, any efforts to re-establish a standards process are doomed to fail.
I've worked with this model for ten years already and there are advantages and disadvantages. Unfortunately, it's rare that you can find either a management team or a software development organziation that understands either.
The component model works reasonably well for the Generic Core Business Functions (i.e accounting, human resources, sales, etc.), however only to the exent that you can use Off The Shelf Functionality. If your requirements are unique (or you think they are) you end up with a full blown development project on your hands anyway (with associated headaches and expense).
If, on the other hand, the requirements you are trying to address are part of the Value-Added Function that your organization provides to it's customers, you had better be willing to invest in creating some real value for the customer. If you spend a couple of months integrating off the shelf components that can provide the same value, you're not likely to be in business long.
Killing projects early for the Right Reasons is simply good management but the need to do this is usually an indication that there is something wrong in the organization itself. If you are part of an organization that does this frequently, the governance process or the development process is broken or both. Flee with all due speed.
What the quote should have read is...
"'You can improve productivity by 20%', Hoyle advised, 'by killing management consultants when you should: which is early in the lifecycle.'"
We've been trying to maintain a product developed in-house over the last decade or so. Wouldn't it make sense to buy a GUI toolkit, they thought, so we can concentrate on our core competency? Sure, except we had to stay on Solaris 8 when everybody else was using Solaris 9, and then 10. The company that provided the toolkit got sold a couple of times, and is now part of some consulting outfit you've never heard of. They have two guys in Bombay trying to port it to newer platform versions, but they don't really test it, so we've had to take on that additional burden. Without the source code. Sometimes they're busy working on other stuff, so they don't get to our complaints for weeks. We're terrified they'll go out of business before we're able to do a rewrite.
Of course, Oracle stopped concentrating on the Solaris 8 drivers, so when we called for support all we ever heard was "upgrade to Solaris x and install the newest version". Would that solve the problem? Who knows? We can't do it anyway because of the GUI toolkit.
Now we want to move that product to Linux, but the GUI tool in question doesn't work on Linux at all. They're trying to get it working on RHEL 3, while we've just moved our other tools to RHEL 4.
You wan't to make a brittle tool and take the blame when the enterprise can't upgrade the desktop OS because a key component vendor just went out of business? By all means, knock yourself out. You can commiserate with one of our groups that's still running Java 1.1 because a piece they bought from a now-bankrupt vendor won't run on a later version. The more third party vendors you have, the worse it gets, too. You can get circular dependencies that prevent you from upgrading anything without a total rewrite.
Me? I'm not writing everything myself, but I use OSS whenever I can. After the number of times we've been burned in recent years, if you work for my company you'd better have a damn good reason to bring in third-party vendors. We're pretty much down to Java and Oracle as the only easy sells for new projects.
If it takes you 2 min to develop the application, I guess that's not the same kind of application Gartner has in mind. The smallest application I worked on was about 120 Man-Days development (not including QA, Analysis and "corporate crap overhead"). Even if you can code 10 times faster than me, that's still 2 week worth of work and that's not something you just want to throw out the window. ( and even if you have time to kill, you would not want to go through QA, TAT, UAT, xxT, xxT2, ... without a good reason )
...
What gartner has in mind is telling the manager what they already believe. Several year ago it was so fashion to rewrite an application from scratch. As a manager, saying that you were reusing something made you look so old school, not a true dot-com mentality. Nowadays you must sacrifice a chicken to get some hope of having the budget to look at the code.
Look at the buzzword friendly tech in the development world, like SOA and Co, this is all about flow management, gluing application together,
I don't know what gartner is for. Basically whatever is the tendency of the day, they just acknowledge that the right way to go.
We are living a land where managers are rewarded for cutting jobs, then rewarded again for selling off units that are failing because they don't have anybody working for them.
This is another picture of the ugly underside of a capital-driven economy. Management reacts to the behavior of a bunch of coked-up day traders and brokers who are shooting craps with other people's money. Because "The Market" likes it when jobs are cut, managers cut jobs. Any way they can dump a few more workers is good, they think.
But the fact is, somebody's got to do the work. The managers certainly aren't going to do it because in the process of getting their MBA all they learned was that if you bought an 8-ball, and sold some to your friends after stepping on it a few times, you could get high for free. So now they're middle management and they don't know fuck-all about actually making something, or providing a service besides oral favors for their bosses.
I know this is heresy in this day and age, but it really is worth the few hours it takes to read Das Kapital. Not that I want to see a Marxist system here (or anywhere, really), but it's worthwhile to know how capitalism looks when it starts to fall apart. And falling apart it is, make no mistake.
Greed has brought us a bloody war in Iraq. It's brought us a middle class whose debt is increasing as they are told they're better off. But the only ones that are better off are the credit card companies.
Come on, let's have a show of hands: Think back half a decade. Think about where you thought you'd be five-years' hence way back then. Have you made it? Are you better off now than you were? Chances are, that unless you are in the very highest levels of management, you are barely scraping by. Sure, you've got an Audi, and a hi-def TV to watch Dancin' with the Stars, and surely your $250 a month cellphone bill is evidence that you're making progress, right?
This year, Americans now have a NEGATIVE savings rate. When it comes right down to it, that's the surest sign of the direction of an economy. How many of you pay off that credit card bill every month and put a little bit aside? Oh, and your retirement account at work doesn't count because that's going to disappear, Enron-style, long before you're able to use it.
No, you really should take a few hours out of your sodden, beaten-down lives and read Das Kapital. And you there. You, shaking your head with the know-it-all smirk. Read it sober.
You are welcome on my lawn.
kill development projects early, "and often," he said, "if your failure rate is high." You can improve productivity by 20%
By killing all of my projects, I'll have a failure rate of 100% but I'll do it 20% faster. Awesome!
Thanks, Gatner.
Friends don't help friends install M$ junk.
But this type of IT organization is not limited to games. I work for a very successful business equipment solution company, and it used to be almost exactly as the GP described. I say Used To because the top brass finally realized that IT was blowing money like crazy. Now they have an 80% turn over rate in 2 years on the network side of the house, 30% on the apps side, budget cut backs, staffing cutbacks, and all round low moral. I recently wrapped up a pair of BSs in IT and Technology Management. I've been working hard to implement project management procedures, and get some kind of order in the shop, but my supervisor is content in code & fix mode, and my manager is 3 years from retirement and has no idea what IT Alignment is. And this has hardly been limited to this company, I have worked for enough organizations to know that this is the more common approach to IT, not the exception. And that is why I'm trying to move from Code Monkey (tm) and Project Management.
-Rick
"Most people in the U.S. wouldn't know they live in a tyrannical state if it walked up and grabbed their junk." - MyFirs
Scope can be reflected in requirements. If the number of requirements goes up 1% per month, you are getting a 1% scope creep. Note that this measure does not take into account differences in complexity between individual requirements - for a more accurate measure one might use function points, number of classes and/or methods needed to implement the functionality, etc. You need a fairly mature process to be able to measure these at all, let alone accurately, but they are available to those who work hard at the process game (now whether or not that actually gets you anywhere is open to debate, but...).
However, the quote of a 1% per month scope creep as a cut off point seems a bit low, especially if taken literally on a month-to-month basis. Across the life of a project, 1% per month may be high, but cutting a project off because its scope has risen 7% in three months of the late analysis or design phases seems a bit excessive, especially if by doing so you have 0% increase during implementation, test, or deployment.
That is all.
I've heard for over 20 years that "soon" there won't be a need for programmers, since you will just be able to make applications by putting components together.
This appeals to managers, since having programmers around is a pain. They're expensive, unreliable and act weird.
It will of course never work in reality. Anything interesting enough to be worth doing will be hard enough to get right that you need an actual programmer for it. And programmers have been putting components together almost since the dawn of programming, it's just that some people don't know that.
Buhahahahhahahhahhahhahaahahahah!!!!! HERE THEY GO AGAIN!!!!!
I'm writing software for 15 years. And I already lost count of how many times I was told that crap and that "tomorrow" I would be among unemployed, since even idiots would be able to create software using tomorrow's modular platforms. That tomorrow is yet to realize.
I was working with asm/pascal 15 years ago - and were basically rewriting applications from scratch: for new platforms and for new performance requirements.
I now work with C/C++ - and basically rewrite applications from scratch: for new platforms and for new performance requirements.
What'd changed? NOTHING.
Would the idiots ever learn? As computer industry develops and grows - so do requirements for computers. 15 years ago nobody expected to have affordable real-time 3D graphics or on-line simulation algorithms or real-time video encoding. Now we take that for granted. As old fart, of course, I cannot even imagine what would be capable computers in next 15 years - but all that would be possible because of abundance of cheap HW performance and I hope more intelligent software. Not because we would have such performance - but because I'm sure there is and would be ever growing demand for it.
e.g. some AI guys might tomorrow implement autonomous OS which would be voice/etc controlled. So you would be able to plug your photo camera into computer, say "Grab all new photos" and (miracle!) it would do that. Then say "If there are more than N gigs of new photos burn me them on disk as photo album.". Etc. Voice recognition + intelligent interpretation of commands + AI personalization - are tasks not yet possible for computers both hardware-wise and software-wise. NOW. How well fit modern algorithms and applications for tasks in such environment? They are completely unfit. So when development would come to that point - we software developers would have to rewrite all the components and blocks to fit well into new platform/OS/etc to get most out of it. IOW, do not put your GCC aside just yet.
All hope abandon ye who enter here.
There are two lines of thought to answer "Why would you ever code an app from scratch again?" - one is simply that if you're doing something worthwhile and innovative, there IS NO OFF THE SHELF solution. If you can buy the solution, is your app doing anything that isn't already being done? The other is tying yourself to a vendor. I know anything from Gartner is designed to promote this, but in the real world, do you want your killer app at the mercy of a vendor who can drop support, go out of business, change the license model, etc - it's not a good business case to put your product at the mercy of third parties, especially now. (Especially with the quality of code in some instances. It takes less time to do it right yourself than buy-rather-than-build in some cases. Not all.)