ThatObviousUser
ὠαἴÏÏιÏÏÎ¿Ï Ïαá¿Ï εἶ
You're just jealous web programming is so much more fun.
-
Hey Guest. Check out your NeoGAF Wrapped 2025 results here!
GAF Devs Thread - a thread about programmer's rants
- Thread starter nickcv
- Start date
- Status
You're just jealous web programming is so much more fun.
The amount of code I have to go through just to draw text on the screen using DirectX11 is ridiculous. And I have to preinitialize all the text, so I can't just tell my graphics class to say "Hello World" whenever I want to, but I need to create a sentence object which is passed through the text engine which is called from graphics, all the while passing back and forth the device context like a blunt in a drug den.
Fuck graphics programming ;-;
Fuck graphics programming ;-;
The Technomancer
card-carrying scientician
Graphics programming still blows my mind to this day. Even something as simple as straight ray-tracing is still being computed for every pixel on the screen sixty times a second. Blows my fucking mind.
Axalon
Member
I work with this guy, where whenever I correct something he does or suggest he do something different, has to argue every little thing about how it's not necessary. I was telling him that when developing for mobile you want to be extra careful about writing valid markup, since mobile browsers are more strict about that sort of thing, and his response was something to the effect of "well it doesn't really have to be valid markup, it just has to work right".
I WANT TO PUNCH THINGS GAF
I WANT TO PUNCH THINGS GAF
A position opened up for a "LAMP - Software Developer II" at my current workplace. I looked at the requirements* and after considering my own (lack of) qualifications**, I decided to submit an application. I re-did my whole resume to reflect the few solid projects I'd worked on that could reflect on my skill set, put in a few good references that could vouch for me, and wrote a nice, slightly ass-kissing cover letter.
The hiring manager happens to be my current manager, so immediately after walking over to HR and dropping off a copy of my application to them, I walked into his office and gave him a copy. We had a chat, and uh, he was surprised, but I assured him I felt I could do the job. With a bit of a raised eyebrow, he warned "you know, we will test you right," and I played it cool and said "of course."
This was on a Friday, so of course my weekend was spent feverishly going over design practices, algorithms, and all the stuff I was always told would be expected of me in an interview in this field.
Monday comes, and I'm given an assessment to take on the spot, which is full of platform/architecture specific questions. The fun part was not remembering much of anything of specific Ruby syntax on the spot, so I ended up just writing a solid paragraph explaining my answer without referencing any specific calls whatsoever.
Then, I was given a project to complete. Two-ish hours and 400-ish lines in Sublime Text later, I turned that in.
Throughout the whole process, I remained calm by basically reminding myself that the worst that could happen would be me tanking the interview/assessment, and I'd end up continuing in the same not-too-horrible position I was in. At the very least, I'd learn something and it might be useful in a future job application.
And then Friday, I get called into my manager's office, and he tells me that his lead went over all the final applicants (apparently there were 5) assessments and projects, and his recommendation based on quality of work alone (he wasn't there for the interview) was to hire me or this one other applicant immediately. Apparently, the word used was "superstar". So, at this point, I'm surprised/happy, but the pessimist in me is already thinking "he's letting me down easy," and then he finishes up by saying "given two equal options, I'll take the guy I've known for four years."
I got the job!
Now, I wrote all this for two reasons:
1. Hell yes, I'm bragging.
2. A non-insignificant part of me is absolutely terrified.
Software development in general has been a career goal of mine for years, and I've been doing it, studying it in some capacity for well on fifteen years, but this will be the first job where coding is my #1 responsibility.
I've been lurking this thread since it started last year, and this is my way of saying hi, and I hope I get to become one of the guys who complains about the stupid co-workers, and not the guy so many of you complain about.
Hello Dev-GAF.
* B.S. in CS, 5+ years experience working in MySQL, PHP, Perl and ideally some experience with Ruby
** A.A. degree, very little documented work experience, but 5-7+ years of consistent experience working with MySQL/PHP/Perl on personal projects (both solo and in teams), and a basic understanding of Ruby/Rails syntax/methodology
The hiring manager happens to be my current manager, so immediately after walking over to HR and dropping off a copy of my application to them, I walked into his office and gave him a copy. We had a chat, and uh, he was surprised, but I assured him I felt I could do the job. With a bit of a raised eyebrow, he warned "you know, we will test you right," and I played it cool and said "of course."
This was on a Friday, so of course my weekend was spent feverishly going over design practices, algorithms, and all the stuff I was always told would be expected of me in an interview in this field.
Monday comes, and I'm given an assessment to take on the spot, which is full of platform/architecture specific questions. The fun part was not remembering much of anything of specific Ruby syntax on the spot, so I ended up just writing a solid paragraph explaining my answer without referencing any specific calls whatsoever.
Then, I was given a project to complete. Two-ish hours and 400-ish lines in Sublime Text later, I turned that in.
Throughout the whole process, I remained calm by basically reminding myself that the worst that could happen would be me tanking the interview/assessment, and I'd end up continuing in the same not-too-horrible position I was in. At the very least, I'd learn something and it might be useful in a future job application.
And then Friday, I get called into my manager's office, and he tells me that his lead went over all the final applicants (apparently there were 5) assessments and projects, and his recommendation based on quality of work alone (he wasn't there for the interview) was to hire me or this one other applicant immediately. Apparently, the word used was "superstar". So, at this point, I'm surprised/happy, but the pessimist in me is already thinking "he's letting me down easy," and then he finishes up by saying "given two equal options, I'll take the guy I've known for four years."
I got the job!
Now, I wrote all this for two reasons:
1. Hell yes, I'm bragging.
2. A non-insignificant part of me is absolutely terrified.
Software development in general has been a career goal of mine for years, and I've been doing it, studying it in some capacity for well on fifteen years, but this will be the first job where coding is my #1 responsibility.
I've been lurking this thread since it started last year, and this is my way of saying hi, and I hope I get to become one of the guys who complains about the stupid co-workers, and not the guy so many of you complain about.
Hello Dev-GAF.
* B.S. in CS, 5+ years experience working in MySQL, PHP, Perl and ideally some experience with Ruby
** A.A. degree, very little documented work experience, but 5-7+ years of consistent experience working with MySQL/PHP/Perl on personal projects (both solo and in teams), and a basic understanding of Ruby/Rails syntax/methodology
Congrats BreakyBoy, welcome to our hell!
As to me, I don't know if I'll make sense here but I'm frustrated because of where I am in my career progression. I'm currently an Senior Solution Architect as such I not only deal with project management and meeting with business and leadership and all that but I also code.
So I get calls about jobs and it always goes like this:
"We need a developer! You want HOW much?!"
"Well I'm more than a developer. I meet with executives and set strategy and all that."
"We just need someone to code."
or
"We need someone to be project manager. You're a developer."
"Yes I can code but I also do all that other stuff."
"Sorry, we don't need a developer."
Argh, the career progression for what we do is really screwed up.
As to me, I don't know if I'll make sense here but I'm frustrated because of where I am in my career progression. I'm currently an Senior Solution Architect as such I not only deal with project management and meeting with business and leadership and all that but I also code.
So I get calls about jobs and it always goes like this:
"We need a developer! You want HOW much?!"
"Well I'm more than a developer. I meet with executives and set strategy and all that."
"We just need someone to code."
or
"We need someone to be project manager. You're a developer."
"Yes I can code but I also do all that other stuff."
"Sorry, we don't need a developer."
Argh, the career progression for what we do is really screwed up.
No. Tabs are adjustable across different editors - and even in different instances of the same editor. So, if you like 4 space indenting, but your colleagues all prefer 8 spaces then using tabs means you can all be happy... you can make their code look like yours and they can make yours like theirs just by changing the tab size. If you use spaces everyone loses.
Edit: ok, so i just replied to something that was posted 6 pages back. Sue me. Spaces annoy me that much
Your position sounds like mine. I don't even know what my official title is anymore but I'm commonly referred to as a "Tech Lead". I was thrilled when my last project got put on hold and another project needed my skills in development to actually just code to help get something done faster. It was such a nice change of pace to not deal with any of the lead like responsibilities and just code all day long. I miss those days...
Real commit messages some of my coworkers have used to describe their changes. These are just the ones I saw right now when I pulled, not cherry picked. They are indicative of their commit messages overall (well, from 2 of our developers). The commits include changes to at least 5 files.
1) changed the way comparer compares.
2) The take on the collection is no needed anymore.
1) changed the way comparer compares.
2) The take on the collection is no needed anymore.
Randolph Freelander
Member
True story. Wish I was making this up.
Dev guy is writing an integration test. Basically repeats the logic of the code under test right there in the test. Literally copies and pastes the logic right into the test, making the test completely pointless. So we're like... rewrite that mess. Actually test something.
This test involves staging some data, then asserting that the output is what you expect it to be considering that data. Basically, the data makes the extremely simple logic go left or right and produce a different output. Heads or tails. A or B. Easy.
He stages the data, then fires off the code under test, then... fires off code that has nothing at all to do with the code under test and asserts that he gets back what he staged. No asserts having anything to do with the code under test. Nothing. So I'm like... rewrite that mess. Actually test what you are supposed to test.
He then calls me over. Couldn't figure out how to write the test. And I'm like... dude. This is easy. No, really. It is. What's the code doing? And he's all.... "I don't know." Keep in mind, I have this stupid code practically memorized, and it's not even my code. I have looked at it once days ago and know precisely how to test it. This dude wrote the code and has been staring at it for days. No idea what it even does, apparently. I'm like "what is your expectation? Assert against the expectation." And he's like "uhhh..." And so I'm all "ok, look at the code. What is it doing?" And he's like... staring... staring... and I'm like "it would probably help if you scrolled a bit to actually see the fork in the road." (It was a long single line in the relevant bit of code, and he had other stuff on his screen.) And he does. And I'm like "look back at me and tell me what the code does." And he's like "uhhhh...."
So finally, as is the norm, I have to spell out the test for him.
And come to think of it, the exchange was actually worse than that. Because he kept telling me that the expectation was the staged data. And I kept saying "no." The data affects the output, but it isn't in the output. And I ask again, what is the expectation. And he tells me, again, that it is the staged data. Keep in mind, this is his frickin code.
This is normally the point where I walk away, send an email that is just a picture of jar jar binks to our (former) team lead with the simple caption "my give up."
Dev guy is writing an integration test. Basically repeats the logic of the code under test right there in the test. Literally copies and pastes the logic right into the test, making the test completely pointless. So we're like... rewrite that mess. Actually test something.
This test involves staging some data, then asserting that the output is what you expect it to be considering that data. Basically, the data makes the extremely simple logic go left or right and produce a different output. Heads or tails. A or B. Easy.
He stages the data, then fires off the code under test, then... fires off code that has nothing at all to do with the code under test and asserts that he gets back what he staged. No asserts having anything to do with the code under test. Nothing. So I'm like... rewrite that mess. Actually test what you are supposed to test.
He then calls me over. Couldn't figure out how to write the test. And I'm like... dude. This is easy. No, really. It is. What's the code doing? And he's all.... "I don't know." Keep in mind, I have this stupid code practically memorized, and it's not even my code. I have looked at it once days ago and know precisely how to test it. This dude wrote the code and has been staring at it for days. No idea what it even does, apparently. I'm like "what is your expectation? Assert against the expectation." And he's like "uhhh..." And so I'm all "ok, look at the code. What is it doing?" And he's like... staring... staring... and I'm like "it would probably help if you scrolled a bit to actually see the fork in the road." (It was a long single line in the relevant bit of code, and he had other stuff on his screen.) And he does. And I'm like "look back at me and tell me what the code does." And he's like "uhhhh...."
So finally, as is the norm, I have to spell out the test for him.
And come to think of it, the exchange was actually worse than that. Because he kept telling me that the expectation was the staged data. And I kept saying "no." The data affects the output, but it isn't in the output. And I ask again, what is the expectation. And he tells me, again, that it is the staged data. Keep in mind, this is his frickin code.
This is normally the point where I walk away, send an email that is just a picture of jar jar binks to our (former) team lead with the simple caption "my give up."
battle for naboo, you know
Any commit message is better than needing to sift through the diffs to try and determine what exactly was changed or added. And I'll be sure to add tests to the next project I start, wish I would have found out about them sooner!
My favorite: "Updated with latest changes", or some variant thereof.
We just started unit testing a few months ago when we started a new project, and only half the team bothered to learn. We're not doing TDD yet, just writing tests after the fact. It mostly just helps to know when you make a change and it breaks things you didn't expect, or to find bugs by forcing you to examine all your assumptions and test all the expectations. TDD would be a lot more valuable.
I just want to say that I've started to use SASS recently, and it's pretty cool.
I spent the whole day refactoring my CSS, which were all separated into different site sections, and recreating them as SCSS. In the end everything is far more organized that it had ever been. I didn't even use any of the more advanced features, but nesting has been very useful.
I spent the whole day refactoring my CSS, which were all separated into different site sections, and recreating them as SCSS. In the end everything is far more organized that it had ever been. I didn't even use any of the more advanced features, but nesting has been very useful.
BomberMouse
Member
Sass is really good, have you also tried less?
BomberMouse
Member
I haven't, I'm not a fan of doing things client side but they do look very similiar and less seems to be more popular nowdays (thanks to twitter bootstrap)
talisayNon
Member
quick question i'm trying to access with javascript a ListViewItem name in a ListView so I can call the Selection_index changed event handler from client side. any idea how to do that?
C#/ASP.Net
C#/ASP.Net
I wouldn't want my site to look empty if JS isn't enabled, though.
Coworker: "Hey, the documentation you wrote for this piece isn't working."
Me: "What do you mean?"
Coworker: "I copied and pasted the example usage code and it won't compile."
Me: "What's the error message?"
Coworker: "The type or namespace name ExampleClass could not be found"
Me: "Well, you have to change the name of ExampleClass to the class you're referring to."
Coworker: "That's stupid. You need to write documentation so I can just copy and paste it."
Me: "What do you mean?"
Coworker: "I copied and pasted the example usage code and it won't compile."
Me: "What's the error message?"
Coworker: "The type or namespace name ExampleClass could not be found"
Me: "Well, you have to change the name of ExampleClass to the class you're referring to."
Coworker: "That's stupid. You need to write documentation so I can just copy and paste it."
CoffeeExpress
Member
it has been a while since my last post in here.
things at work have become pretty crazy, i moved to another company which is pretty big (we are a little more then 100 employees) and the projects have become a lot more complex as well.
recently we just started working on a new project and the guys in charge of the application templates are keep changing them, day after day, i swear if we would be working into the same building i'd have kicked his ass for the whole day.
speaking about something else recently i just deployed a personal project and after the initial excitement i feel like i've build a cathedral in the desert: no one who may actually belong to my target userbase knows the project even exists, and everything i tried to do to spread the word isn't working in any meaningful way. The worst thing about it is that at this point the only thing i would want is some kind of feedback, just to don't feel like i've wasted months working during weekends and at night.
things at work have become pretty crazy, i moved to another company which is pretty big (we are a little more then 100 employees) and the projects have become a lot more complex as well.
recently we just started working on a new project and the guys in charge of the application templates are keep changing them, day after day, i swear if we would be working into the same building i'd have kicked his ass for the whole day.
speaking about something else recently i just deployed a personal project and after the initial excitement i feel like i've build a cathedral in the desert: no one who may actually belong to my target userbase knows the project even exists, and everything i tried to do to spread the word isn't working in any meaningful way. The worst thing about it is that at this point the only thing i would want is some kind of feedback, just to don't feel like i've wasted months working during weekends and at night.
never had to deal with it, but i had the pleasure to work with the facebook api: the worst thing i've ever dealt with in my whole life.
The SQL editor of MS Access is a fucking mess.
There's neither line numbering nor syntax highlighting, the controls for changing the font and font size are hidden deeply somewhere in the Access settings, and strg+a does not work at all.
It occasionally puts table and column names between square brackets for some reason (e.g. [emp].[id]) and automatically adds column names for the results in the local language, e.g. in my case AnzahlVonId for count(id).
If I temporarily create a new query for testing something, then open another, it's not possible to switch to the unsaved one again. Unsaved queries are not accessible from nowhere in the interface, you have to save it first.
And the worst of it: If you use an unspecified identifier in a query, Access shows a dialogue prompting for its value when the query is executed, which can be pretty convenient. I was working on a complex query, when at some point the parameter prompt appeared - but the query contained no parameters. I fiddled around, rewrote and simplified the query several times, but it kept asking. Boggled my mind.
During testing, once I had sorted the query results in the result view. Turned out Access had added this sort key to the metadata of the query. Later I didn't use that column anymore, so Access didn't know it and thought it was a parameter. That's fucking bullshit.
There's neither line numbering nor syntax highlighting, the controls for changing the font and font size are hidden deeply somewhere in the Access settings, and strg+a does not work at all.
It occasionally puts table and column names between square brackets for some reason (e.g. [emp].[id]) and automatically adds column names for the results in the local language, e.g. in my case AnzahlVonId for count(id).
If I temporarily create a new query for testing something, then open another, it's not possible to switch to the unsaved one again. Unsaved queries are not accessible from nowhere in the interface, you have to save it first.
And the worst of it: If you use an unspecified identifier in a query, Access shows a dialogue prompting for its value when the query is executed, which can be pretty convenient. I was working on a complex query, when at some point the parameter prompt appeared - but the query contained no parameters. I fiddled around, rewrote and simplified the query several times, but it kept asking. Boggled my mind.
During testing, once I had sorted the query results in the result view. Turned out Access had added this sort key to the metadata of the query. Later I didn't use that column anymore, so Access didn't know it and thought it was a parameter. That's fucking bullshit.
you win XD
Like using VBA?
I'm currently working on a VBA script to create some fancy charts in Excel out of the Access data. I had to create a percent-based bar chart but the bars didn't necessarily have to add up to 100, so I just added another row with the value totalSum-sumOfOtherRows and intended to make the respective part of the bar transparent and remove the entry in the legend. The first part went well, but as for the legend - the LegendEntry objects do not contain their title. The title's stored in the SeriesCollection which also stores the bar part values etc. They're ordered the same, so you just have to look for the title in SeriesCollection and remove the item with the same index from LegendEntries.
Well, nope. It's reversed. The order of the LegendEntries is fucking reversed. What the hell? And hooray for compatibility: They fixed that with Office 2007.
Yes. Everything is in VBA. Everything is Access. Both the front and back ends. For some reason I'm not allowed to develop applications in anything other than Access per the IT department head. And this isn't just some "click a button generate a chart" type of stuff either. This is production critical "everything shuts down if one piece isn't working" type stuff. As in "used by an entire warehouse worth of people when Access databases aren't really meant to be used by more than a handful of concurrent users" type stuff. The pain I have to deal with because of versioning differences, or limiting functionality, simple stuff like generating PDFs or exporting reports as Excel files because of the licensing issues or whatever was going on when Office 2007 came out and that some people don't have the service pack that addressed that, or the emails I get that people can't access data because someone doesn't have the correct permissions to the network folder where the backend is located, or having to create a macro for a basic function like right-clicking a datasheet because it isn't supported in the Runtime version. Or the fact that one of the few Windows apps I did get to build chugs because it connects to Access using a .NET layer on top of a JET layer. Because I had to convert a perfectly good SQL Server database into an Access database, because, one, the server was my workstation (that I had to leave on and locked otherwise nothing worked) and I couldn't get a server space to work with until about a month or two ago, and two, Access only! The only reason the fron end wasn't converted is because I just never got around to it, and the IT department at least has the understanding that doing so would be a detriment to production. Or just the general feeling that everything in Access is held together by silly string. Parameterized queries? What are those? Distinct count? What's that? Asynchronus work? Never heard of it. Features and datatypes that are often advised that you straight up don't use because of how buggy they are. Random and undetectable corruption, on the front end, you know, the part that DOESN'T house data. Working in design view gets slower and slower as the scale of the application increases. Doing things like clicking on Design View from Form View, setting a Record Source, or even something as simple as maximizing or restoring a window straight up hangs Access, sometimes to the point where it becomes a big white screen. Oh, and God forbid my computer freezes or the power goes out while I'm the middle of something. There's a good chance that everything is screwed.
I mean, it's cool that I now have some strong VBA skills, but I haven't had much of a chance to hone or improve my skills using "real" languages since college, and it feels like what I do know is eroding, and it's hurting my chances of ever being able to further my career. It's not a cost thing either. I have Visual Studio and SQL Server sitting right here on a virtual PC to maintain the few applications I developed before the IT department came up with whatever asinine reason that I'm just flat out not allowed to develop with it. It's a bit disheartening, especially considering that I could have a smoother experience for myself and users by building Windows apps that use SQL Server databases. I'm honestly starting to get less motivated and feel more unfulfilled by the day. I've been trying to counter this by brushing up on C# and trying to teach myself HTML/CSS/Javascript and ASP.Net in my spare time. Maybe see if I can get some side projects as learning/refreshing experience, especially seeing how a lot of the dedicated software development positions want web application experience/knowledge.
Heh, this became a full on rant in reponse to a three-word question.
Another thing about the Access SQL editor.. If you change the font in the settings, the new font/size isn't applied immediately, only for new queries or if you open an existing one for the first time. If you have already opened one, well TOO BAD. You can't close it. You have to close and reopen the whole Access file.
The same applies for renaming a table - it's not possible once you've already opened it (= viewed its contents).
The same applies for renaming a table - it's not possible once you've already opened it (= viewed its contents).
Question for those of you who manage code projects.
I tend to overload my schedule through design and at the beginning of test and even have an informal test period before it even gets to test. I do this because the idea is that it will allow my developers to find the glaring bugs early and fix them before we get to test.
This usually ends up meaning that my devs work like mad through design and then again right up until test. And then the rigor is usually so high on the informal test cycle 0 that they end up not having much to do and take a bit of a breather once we hit actual test.
I've been hailed as a hero and critcized as a 'time waster' for this. In that it means you come to our cubes during test most of my developers are sitting waiting (I tend to roll off developers to other projects once we get to test, but you gotta keep at least a few for new bugs). So I've gotten feedback from some that I'm not pushing hard enough or from people that understand how hard I push my devs during design/before test they think it's wrong to overload early to have a drama free test and go live they'd rather me keep a lower burn throughout the whole thing.
But I've found that never works. You keep a lower burn it'll blow up in your face once you get to test and everyone's running around like mad.
What are your thoughts? I tend to like the whole "very thorough design, even more thorough unit test. Easy almost wasteful test cycle." approach versus the whole 'ok design, ok unit test, possibly very dramatic test cycle."
I tend to overload my schedule through design and at the beginning of test and even have an informal test period before it even gets to test. I do this because the idea is that it will allow my developers to find the glaring bugs early and fix them before we get to test.
This usually ends up meaning that my devs work like mad through design and then again right up until test. And then the rigor is usually so high on the informal test cycle 0 that they end up not having much to do and take a bit of a breather once we hit actual test.
I've been hailed as a hero and critcized as a 'time waster' for this. In that it means you come to our cubes during test most of my developers are sitting waiting (I tend to roll off developers to other projects once we get to test, but you gotta keep at least a few for new bugs). So I've gotten feedback from some that I'm not pushing hard enough or from people that understand how hard I push my devs during design/before test they think it's wrong to overload early to have a drama free test and go live they'd rather me keep a lower burn throughout the whole thing.
But I've found that never works. You keep a lower burn it'll blow up in your face once you get to test and everyone's running around like mad.
What are your thoughts? I tend to like the whole "very thorough design, even more thorough unit test. Easy almost wasteful test cycle." approach versus the whole 'ok design, ok unit test, possibly very dramatic test cycle."
captainblye
Neo Member
....Are.....Are you me?
i like this approach.
qa phase is about fixing bugs, which i hope is isolated to just that. none of this, "oh, to fix this bug, it's gonna approach a feature-level work because we didn't think of this case during the design." meanwhile, other bugs continue to pour in. plus, there's great satisfaction in writing bug-less code the first time around.
Phoenix
Member
This is the approach I espouse. Burn in the team early. Gets through the Tuckman group model VERY quickly. You CAN burn out a team too early following this approach, but its better than the death march you incur if you don't do this IMO.
Good to hear that others do this.
I often get the complaint "The developers are just sitting around!!"
"Well are we in test?"
"Yeah."
"And are their any bugs?"
"No."
"Soooo?"
I much prefer the early burn instead of the sustained madness. Honestly people complain about my devs not being so busy during test but they don't really pay attention when I'm working them hard at Design and Unit test.
Also it makes me laugh when I make plans to roll off some developers to another project during test and people are like "WHAT?! ARE YOU MAD? KEEP THEM ON." Then later during test when there's a low number of bugs they're like "Why aren't they doing anything?!" Sigh.
What's a test cycle? I'm not a manager, but as a developer I just do a quick smoke test whenever I complete a feature, then roll a package off to the test team. They do rigorous testing against the feature, give feedback on the UX, and some basic testing throughout the software. Automated unit tests and integration tests catch basic regressions, though manual testing has to be done to catch the less obvious ones.
Bugs are fixed as they're found, and releases are pretty much invisible to developers. We put out a releasable package every 2 weeks. Finishing all planned features just means moving to a new project. I occasionally get a day or two of free time if I finish a feature early and can't meaningfully contribute to another feature, but that time can be spent on improvements to my development environment, the code base or our tools.
Yeah I guess methodology has a lot to do it.
I should've pointed out on projects with a more traditional waterfall methodology with long drawn out test cycles.
Do you have to work that way? What do testers do when you're not in test?
Depends on the client. Usually the testers have other jobs or in the case where we don't have enough the contract out the testing.
RiccochetJ
Gold Member
Eh, the only problem I ever really have is when our new hires discover cursors.
ahah reminds of this one time a new hire was complaining about his cursor hanging.
I looked at it
cursor c1 is
select SYSDATE from dual;
For c1_rec in c1
Loop
...
I felt bad later cause I laughed so hard I nearly cried.
Another Access story.
We've got a main person table and another one with a lot of, let's call them processes. Every person has 0 to n processes. Also, there are 5 special processes which are part of the person records, as everyone can maximally one of them.
I had to calculate several statistics about all processes, so I created a view containing several unions.
The schema was: person id, process id (if it's from the process table, else null), process description (not unique) and additional data.
Now, the tables are stored in an Oracle database and just referenced by Access. Everything seemed to work well, until I noticed a duplicate entry when opening the view in Access - which should have been impossible. I reopened it, and it showed a different duplicate entry. What the hell? I had a look at the query. No errors. I executed the query in Oracle - no duplicates.
I attempted to re-add the view to Access. If there's no primary key constraint, Access asks which columns should be regarded as primary, if any. That was when I realized what was wrong: When I linked the view in the first place, I just set the first two columns to primary without thinking much about it. I thought that's probably just for avoiding integrity violations, but it was a non-updateable view anyway, so whatever. Turned out Access uses this information to decide which data to pull from the database. For each set of primary keys, it returns one record. As I mentioned above, the special processes from the person table had null values instead of an id, so basically for every person with more than one special process, Access always pulled the same record for each special process. Because it couldn't distinguish them.
I hope I didn't explain this in a too confusing way.
We've got a main person table and another one with a lot of, let's call them processes. Every person has 0 to n processes. Also, there are 5 special processes which are part of the person records, as everyone can maximally one of them.
Unnecessary null values, I know. I didn't create it and I've got no time to revise the data model.
The schema was: person id, process id (if it's from the process table, else null), process description (not unique) and additional data.
Now, the tables are stored in an Oracle database and just referenced by Access. Everything seemed to work well, until I noticed a duplicate entry when opening the view in Access - which should have been impossible. I reopened it, and it showed a different duplicate entry. What the hell? I had a look at the query. No errors. I executed the query in Oracle - no duplicates.
I attempted to re-add the view to Access. If there's no primary key constraint, Access asks which columns should be regarded as primary, if any. That was when I realized what was wrong: When I linked the view in the first place, I just set the first two columns to primary without thinking much about it. I thought that's probably just for avoiding integrity violations, but it was a non-updateable view anyway, so whatever. Turned out Access uses this information to decide which data to pull from the database. For each set of primary keys, it returns one record. As I mentioned above, the special processes from the person table had null values instead of an id, so basically for every person with more than one special process, Access always pulled the same record for each special process. Because it couldn't distinguish them.
I hope I didn't explain this in a too confusing way.
- Status