|
A GOOD BA is invaluable, a mediocre one causes all the problem you have described. I always managed to get myself included in the initial requirement studies, the follow up crap was left to the BA.
Never underestimate the power of human stupidity -
RAH
I'm old. I know stuff - JSOP
|
|
|
|
|
Mycroft Holmes wrote: A GOOD BA is invaluable, a mediocre one causes all the problem you have described. Yes, we have both good and bad - and, (maybe not that clear from my OP), I wasn't saying that all BAs do a bad job. I was questioning their necessity. When working in a software development role, I have always preferred direct and continuous interaction with the end user. I find it much easier to get a proper understanding of their requirement and I, therefore, have a much improved chance of delivering the right solution, with fewer iterations.
|
|
|
|
|
If you’re short on staff I would say no as a good dev will cover the analysis. This is where I like to be.
If you’ve got the staff then I see BAs as a specialism: developers can spend more time on the technicalities whilst the BAs can dig deep on the “whys” around specific business practices, unravelling complex processes that as a developer you might find tedious. However, BAs and Devs need to both attend the core meetings to that BAs don’t become a go-between or worst, a poor information conductor
|
|
|
|
|
|
I'm not sure that you work on problems quite as complicated as I do.
I work about 10 hours a day. Frankly, I need all the help I can get. I need the help of DBAs that know the connections and flow between the ... 20 or so databases that my system interacts with. There are literally thousands of sprocs. I have no idea what is in most of them. It's the same way with BAs. I've had useless business analysts and I've had ones that made my job a joy. Dealing with customers can be time consuming and problematic. If someone can offload that task, that's great.
|
|
|
|
|
Michael Breeden wrote: I'm not sure that you work on problems quite as complicated as I do. Over 50 front-end web applications, (some with their own databases), all interacting with our legacy system, which consist of over 1000 tables and 10,000 pieces of software. I wouldn't call it simple.
Note. I use the term 'legacy' to describe it's long standing value - not as a derogatory term commonly used in IT these days.
|
|
|
|
|
Absolutely agree with all of the comments. I found that BA's actually make things worse because they do not understand enough about the coding end to prevent business users from creating a mess.
Realistically, a total waste that creates even more work. 'BA' = Business Agent because there is never any true 'analysis' being done, simply a pass-through person or go-between.
|
|
|
|
|
Member 14840496 wrote: Realistically, a total waste that creates even more work. 'BA' = Business Agent because there is never any true 'analysis' being done, simply a pass-through person or go-between. Yep! That's where we are at. I've seen comments from developers, on here, saying that it saves them time. But I prefer to spend time, fully understanding the problem by talking with the person who has the problem - rather than spending my time coding something that the BA (a) didn't fully understand and (b) didn't properly communicate the bits that they did understand. Chinese whispers comes to mind!
|
|
|
|
|
I believe the BA's do serve an important part of the business, but like someone else mentioned on here, they need to have domain experience or working to gain that experience.
Ideally, as software engineers, we are all very, very busy (thankfully!). When the business wins a new customer or an enhancement for an existing customer, the BA's come in to clarify the customer's needs from a "business" perspective.
Meanwhile, us software engineers are working diligently to keep up with the work we already have so it's nice that we have the BA's out there getting our future work ready.
When the new work is defined enough from a business standpoint the engineers can be brought in to start figuring out the technical details to satisfy the business requirements. The BA's are consulted about requirements as the tech team comes up with a plan.
As the project moves forward the BA's will start to focus on other needs since they won't be consulted as much as they were in the early stages.
I don't know what company your with, but I'll bet you are just seeing growing pains since the BA role is new to the company. The BA's are probably still learning the domain so they aren't as effective as they will be in the future.
Hang tight if you like the company. The BA's should make your life better once they get some runway.
|
|
|
|
|
I have to agree with you. I've never worked with a BA who was worth their salary, but I'm sure there are some out there.
My Agile story is when "we're going to try Agile on this project!". Then everyone except the technical people go away, and come back six weeks later with a 60+ page requirements document written by the BA.
Sigh. No, that's the opposite of Agile.
Of course, the document was so poorly considered, we had to do agile-like cycles of development/review/etc anyway. Only six weeks later.
|
|
|
|
|
The (flawed) business reasoning behind the proliferation of PM's and BAs is that because they work for less money, overall cost can be lowered for development by using less (and more expensive) software developer/engineer positions. The (again, flawed) thinking is that these non-technical PMs and BAs are offloading on a one-for-one person-hour the PM/BA type work from the developer.
Of course, this (flawed) thinking comes from the business types who have no clue how software engineering actually works.
|
|
|
|
|
Interesting, I hadn't thought of it that way -- the average salary decreases by hiring more resources.
|
|
|
|
|
Yes. Been there, done that!
|
|
|
|
|
The two times I recall being handed a spec I threw it out and wrote something better.
|
|
|
|
|
In my opinion, it works best when you don't have business analysts, project managers, or even programmers / coders. Managing projects and analyzing the business needs is the responsibility of software engineers. You get better end results when the people writing the code are deeply involved in understanding the business needs.
|
|
|
|
|
Member 7799927 wrote: You get better end results when the people writing the code are deeply involved in understanding the business needs. You summed it up perfectly, with this one sentence.
|
|
|
|
|
While BA's can be useful, they should be used appropriately as an adjunct help to the developers.
I wrote this about 1-1/2 years ago, and still believe this is true - from years of experience, having done it, and having seen it in action. The trick is finding the experienced engineers capable of project management and technology management with sufficient people and business skills. There are those out there who can, and within that population, those who will. I have seen on several occasions how non-technical management makes a mess out of technical projects. Some just take longer, cost more, and result in mediocrity at best; some were complete failures.
Soup to Nuts[^]
As for Agile...
Agile Principles from a Traditional American View/[^]
Rethinking the Software Development Life Cycle (SDLC)[^]
|
|
|
|
|
Great articles in your links.
|
|
|
|
|
|
I feel like you're describing an FA instead of a BA.
BA's basically mess around by "identifying" abstract business objects of value, typically for internal reporting to a board or a director. Then they make requests for implementing a business layer that respects and reports on those objects.
This is done to give the board members a false sense of security.
In reality, however, business objects are figments of someone's imagination made concrete, for no clear reason other than to show graphs or metrics. This in itself can create value if your brand narrative relies on fancy graphs and is mostly B2B oriented.. but the exact same result can often be achieved by measuring actual objects, so I fail to see the point of mucking up a code base for imaginary points of interest.
|
|
|
|
|
I certainly would not tell you how to feel. You are entitled to your opinions and feelings, and I respect that.
But in looking back over 40 years in software engineering in several vertical markets, and still being active and full time, so in fact and in theory, I was not describing a Functional Analyst.
At the core of modern software architecture is creating abstract entities (e.g. classes) that may or may not describe real, concrete entities. So your run into business objects being figments of one's imagination does not logically follow.
Business analysts that have little or no domain experience and do not understand how software is actually made are a waste of time. A senior-level software engineer can learn the BA aspects of the SDLC and manage the direct involvement with the business side (both process and people) much easier than a BA can learn the technical side. In my career, for example, I learned the business side and customer-relations side of software engineering quite easily, and breezed through getting an MBA.
My point is that BAs and PMs that do not have the technical expertise and experience in software engineering are a net negative to successfully completing a software project with quality, reliability, sustainability, extensibility, with good performance, on time and on budget. Too often, in my experience and IMHO, the BAs and PMs are the bane of a developer's existence and a barrier to making a good product that developers must waste time trying to overcome.
That said, the best BAs and PMs I have known are those that fall into at least one of these categories:
1 - Recognize their limitations and rely on the senior software engineer(s) to turn their business ideas into software ideas at the requirements and design level, and do not try to manage the architecture, design, development, quality assurance, and deployment aspects of the SDLC.
2 - Are willing to learn, put their ego aside, and take the time to learn about the much more complex world of software engineering.
|
|
|
|
|
While I do understand your take on the matter, and don't mean to undercut it in an unrespectful manner, I am gonna challenge the notion that classes could be based on either real or abstract entities.
I'm 20 years deep in OO-design and the major problems I keep coming across with class-based designs, all originate from classes being based on non-concrete entities. As an iron rule within my own team, I demand that all classes are directly based on either concrete entities or the pre-defined abstraction layers we've all agreed upon (so services and data models are OK, helper classes or business objects are not OK). Over the years, this has been the magic sauce that made my team come out on top with regards to velocity and code quality.
I'm playing with the idea of writing something on the topic eventually, because I think it's a relatively novel POV within our sector. I do like to use a quote from the matrix to highlight the underlying sentiment:
When it comes to abstractions, "The problem is choice".
|
|
|
|
|
This goes back to the 90's.
"My" BA would tell me what he got from the user; then he would ask me to explain it back to him.
I quit that company for lack of a bigger picture.
BA's, in effect, think and act like you work for them.
It was only in wine that he laid down no limit for himself, but he did not allow himself to be confused by it.
― Confucian Analects: Rules of Confucius about his food
|
|
|
|
|
Hard no.
Allocating resources to BA's is a mistake. They do not add value for the end user and they increase the technical cost, while reducing the individual ownership for everyone involved. I've seen enterprises run successfully with and without BA's, and without them there is less churn, more individual responsibility and less sunken costs in reports and metrics for internal use only.
They do not add value and they do not contribute to getting the work done, so why waste the resources. Hire more support and customer training positions instead for a much better ROI.
|
|
|
|
|
Yep. Totally agree.
|
|
|
|
|