Floating Development Schedules

The forum for ranting, raving, complaining and praising
jim
Red Lobster for the seafood lover in you
Posts: 50393
Joined: May 1 06, 2:41 pm

Re: Floating Development Schedules

Post by jim »

planet pujolsian wrote:Any others who want to jump in and defend developers who can't make a deadline they chose? There is no scope creep, no hidden gotchas, etc. All they need to do is take a flash version of our existing and port it to html5, while also having it take advantage of iPad features. I just find it amusing they are the only group within any company that never needs to be accountable.
It's hard. Lot's of unknowns sometimes. Things happen. That's life.

TimeForGuinness
Hall Of Famer
Posts: 20007
Joined: April 18 06, 7:38 pm

Re: Floating Development Schedules

Post by TimeForGuinness »

Maybe the guy who like to talk about his fishing rod wandered down to the software guys and distracted them for a few months.

jim
Red Lobster for the seafood lover in you
Posts: 50393
Joined: May 1 06, 2:41 pm

Re: Floating Development Schedules

Post by jim »

Software development is so intertangled with so many variables we will always be giving ballpark SWAG's when it comes to estimates for large projects. In fact, a SE methodology that I've been following the gospel of for some time now in my projects, pretty much just gives up on the whole pointless exercise. Agile programming techniques break things down into tiny well defined parts, and then you give the estimates for that. It's counter to many cultures - it's counter to our culture here where I work - and I've been getting flak. My Docker's wearing MBA trained boss keeps wanting some god damn MS Project spreadsheet on one large project we are working on, and I refuse. I'm just now starting to tell him we will probably be able to roll something out late 4Q, or most likely 1Q of next year. I just won't go into more specific, and his MBA trained mind is ready to explode. He wants to do a power point presentation for the other pointy heads so bad he is quivering. Presentation with pictures, and numbers, and dates and charts and strategeries and such things. Important stuff apparently.

Sorry, ain't happening because I'm staying rooted in reality and I'm not just going to make some [expletive] up so that you can draw a pretty picture on the computer. I can tell you that the event creation factory will be done in two weeks - that's as far as I'll take it. That's one piece out of about 100.

User avatar
vinsanity
Chili dog truther
Posts: 8836
Joined: July 3 06, 2:19 pm
Location: Indianapolis

Re: Floating Development Schedules

Post by vinsanity »

jim wrote:Agile programming techniques break things down into tiny well defined parts, and then you give the estimates for that. It's counter to many cultures - it's counter to our culture here where I work - and I've been getting flak. My Docker's wearing MBA trained boss keeps wanting some god damn MS Project spreadsheet on one large project we are working on, and I refuse. I'm just now starting to tell him we will probably be able to roll something out late 4Q, or most likely 1Q of next year.
I did agile at the last client I was at for 10 weeks. Loved it. You waste 20-30 minutes a day and about 1 hour every month dealing with the documentation issues instead of weeks on end at the end of every project.

jim
Red Lobster for the seafood lover in you
Posts: 50393
Joined: May 1 06, 2:41 pm

Re: Floating Development Schedules

Post by jim »

vinsanity wrote:
jim wrote:Agile programming techniques break things down into tiny well defined parts, and then you give the estimates for that. It's counter to many cultures - it's counter to our culture here where I work - and I've been getting flak. My Docker's wearing MBA trained boss keeps wanting some god damn MS Project spreadsheet on one large project we are working on, and I refuse. I'm just now starting to tell him we will probably be able to roll something out late 4Q, or most likely 1Q of next year.
I did agile at the last client I was at for 10 weeks. Loved it. You waste 20-30 minutes a day and about 1 hour every month dealing with the documentation issues instead of weeks on end at the end of every project.
It's the only thing that makes sense. We haven't completely embraced every single detail - the pair programming thing we kind of sort of do. Lots more collaboration then before but we aren't following the book completely - our scrums just happen because we all sit close to each other.

But the old waterfall [expletive] - it just doesn't work.

For our existing software in production we do 2-4 releases per month. The second the developers have had a chance to do anything that has any value at all - and it could be as simple as changing the red button to green - we go through our change and release process to get it cut. Rather than wait for gobs of features we do many small incremental releases. Sometimes if the feature is larger and maybe requires it broken up into 2 or 3 smaller pieces for manageability, we will release one of those as we test it. So releases that may not even have any noticeable features added - simply prepping for other releases - may be a candidate.

I've just found this methodology much better on so many levels - for one developers can look and see the last number of the release changing and see progress.

I'm really on board with Agile, unfortunately I'm fighting a pretty strong current where I work and in fact have had to implement some Agile techniques in secret.

User avatar
vinsanity
Chili dog truther
Posts: 8836
Joined: July 3 06, 2:19 pm
Location: Indianapolis

Re: Floating Development Schedules

Post by vinsanity »

jim wrote:For our existing software in production we do 2-4 releases per month. The second the developers have had a chance to do anything that has any value at all - and it could be as simple as changing the red button to green - we go through our change and release process to get it cut. Rather than wait for gobs of features we do many small incremental releases. Sometimes if the feature is larger and maybe requires it broken up into 2 or 3 smaller pieces for manageability, we will release one of those as we test it. So releases that may not even have any noticeable features added - simply prepping for other releases - may be a candidate.
It's kind of funny (to me) because ideally that's how software is supposed to be developed. In small incremental pieces, thoroughly testing each part - at least that's what I'm told by some really start fellas. The whole idea behind unit testing.

So the tell developers, here's this giant chunk. Break it up in to smaller chunks and code that, thoroughly test it, then after 1-12 months of 10 guys developing separate pieces we'll put it all together, give it to a QA team and every time a bug pops up you have to wade through months of changes by multiple parties to fix it. A small bug could end up being a catastrophic one, require weeks of rework and you miss a deadline.

Agile makes the SDLC more like the actual development. 'Here's what we want it to do. Here's step 1. Go do step 1 and we'll release it. Repeat'

User avatar
wart57
just can't quit you.
Posts: 25076
Joined: April 18 06, 4:33 pm
Location: Lost

Re: Floating Development Schedules

Post by wart57 »

um....what?

User avatar
TheoSqua
Next Gen Wart
Posts: 8951
Joined: April 22 06, 6:53 pm
Location: St. Louis
Contact:

Re: Floating Development Schedules

Post by TheoSqua »

yeah, I've found that a lot of times development schedules are a bit out of your control.

For example a Line of Business wanted to password protect some documents on the website to keep competitors from easily accessing the information. Password protecting a directory in IIS takes about 5 minutes, 10 if you forget how to do it and need to google it. I could show my grandma how to do it in 15 minutes.

So i'm asked how long it'll take. I ask how many people will be using it, how many username/passwords they need, etc. Get the information necessary to give an estimate. I tell them 2 weeks because the server is in production and we have some strict security procedures that you have to go through to modify that server.

We set it up on test and let them test it, and they lock out the password. Since lockouts on production require a bit of work, it's decided they should make an individual password for each client. They then tell me that's 500 (about 495 more than the lob initially told me when asking for my estimate) clients. It's decided that's way too many usernames to give access to, so we decide to not use password protection from the server and just create a log-in box on the page. How long will that take? I say about 2 hours programming, even though I felt I could get it done in about 30 minutes.

I create the log-in box. But they don't like how it's styled, so I re-do the css and add a custom button. They don't like how it's hard to tell when you're logged-in or not. So i rewrite it so that the session checks the log-in value and displays if you're logged in or not. Next they want a log-out button on every page, then they want a better error page in case the password is incorrect. In all it takes about 5 hours of coding to get what they want. Now, because we're using a different method of implementing the solution, it has to go through IT-security to be approved to put on the website - which is a 2 week process.

So the line of business is upset that something I said could be done in 2 hours and i felt would take 15 minutes to set up to show them has turned into a 5 week fiasco. But the thing is the decision on how to write it was made on information the business unit gave me that was incorrect. If they had been upfront with their needs, rather than skimming on the details on hopes of getting the solution done faster, it wouldn't have taken 5 weeks. But the blame and complaints aren't on the business unit, it's my fault.

Kinda reminds me of this image:
Image

Arthur Dent
Hall Of Famer
Posts: 12542
Joined: April 25 06, 6:43 pm
Location: Austin

Re: Floating Development Schedules

Post by Arthur Dent »

planet pujolsian wrote:More times than not, the developers themselves are asked what is a realistic time frame for completion
In my experience, more often than not, when engineers or engineering managers are asked to project project completion, they are required to do so under intense pressure from business folks to make best case projections. If everything went perfectly, you might be able to get it done by day X. That is what they want to hear. We always explain that there is a high risk of not meeting that deadline or of producing a flawed product (which in my world means hundreds of thousands of dollars and months of delay) to fix predictable issues. And that doesn't even account for the unpredictable issues, which are more or less guaranteed to occur. But they want to hear the best case scenario, so they can make promises to customers, therefor virtually guaranteeing that they will be pissed off later when the foolish promises cannot be met. As a customer, I'd much prefer to hear a longer delivery time and then be pleasantly surprised if things come in early, than to be promised something quickly and then be denied. If you know that these schedules are consistently going to be missed, why are are you still promising them to customers, and then blaming the problem on the developers?

So, I think you may want to try and understand what issues your developers are facing rather than assuming that it is just a matter of accountability.

User avatar
Bo Hart
Perennial All-Star
Posts: 8770
Joined: April 14 06, 9:50 pm

Re: Floating Development Schedules

Post by Bo Hart »

planet pujolsian wrote:Any others who want to jump in and defend developers who can't make a deadline they chose? There is no scope creep, no hidden gotchas, etc. All they need to do is take a flash version of our existing and port it to html5, while also having it take advantage of iPad features. I just find it amusing they are the only group within any company that never needs to be accountable.
I'm a web developer. I do some design work (as in visual stuff -- the interface of the site, how a user interacts with it, etc), and then a lot of behind-the-scenes work stuff too: interaction with databases, a lot of data processing, writing programs to dynamically draw maps (etc). The scope of web development is pretty broad.

You are completely right that deadlines aren't met very often. When I was an intern here, they had me work on this project that they wanted rolled out in late February. I think we got it out the door in early April or something. The project I'm working on now was supposed to be done in early August, and we're still fixing the last few remaining issues. As much as my boss pushes deadlines, they're always soft. He's never been angry when I've failed to meet a target date.

I think what vinsanity said is right on target. You can get away with blowing by deadlines because its such an unpredictable field of work. A coworker of mine spent upwards of three weeks working on a problem that I legitimately thought would take two days max. He wasn't even writing new software, he was just incorporating an existing feature into a new product we had! In fact, he still has lingering issues on it.

Programming is inherently unpredictable, but web development especially so. You have to design a product that will be accessed by a handful of different browsers. Don't underestimate how difficult that is. You can have it functioning perfectly in one browser and be flat-out unusable in another. There are standards out there that browsers are supposed to conform to, but it just-so-happens that some of them blatantly ignore them (*cough* Internet Explorer *cough*). That just makes testing that much more difficult, complex and time-consuming.

Post Reply