|
comedyblissoption posted:... Work really should just be organized around taking poo poo off a priority queue. this would require management / stakeholders to be able to implement a priority queue, which is where I find most disagreements happen. Every company I've worked for has decided at one point or another that the assigned priority of any given task "doesn't count because we didn't know about X".
|
#
?
Aug 31, 2018 14:32
|
|
|
|
| # ? May 23, 2024 12:19 |
|
|
That's where the PO's need to jump in and shield the Dev team from the squabbling. The PO's need to work with the stakeholders to determine business value, work with the Dev team to identify dependencies, and then get the backlog in the right order. The Devs should never see or care about the "Business Value" field and the stakeholders should never see or care about the "Story Points" field. Naturally, the official Scrum Guide provides no insight on how to get a giant backlog into the right order, so nobody ever really learns how to do it. My boss lamented the disorganized nature of (what we call) our backlog and I offered to stretch into a PO role. They were excited about the idea until I mentioned that a PO basically spends a third of their work hours just farting around with the backlog. So we stuck with our ad-hoc project management by "champions" with not even informal project management training.
|
|
#
?
Aug 31, 2018 15:17
|
|
|
I wanted to post a really long post but thought against it. Anyways I genuinely have really bad 'imposter syndrome' (like, as in, when you are an actual imposter ?? like when your coding skills appear to have gone backwards on the job rather than forwards? except for doing particular, in house custom 'busy work'. The isolation of working remotely has set in really bad I probably should make an appointment to talk to someone even. Murrah fucked around with this message at 15:25 on Sep 1, 2018 |
|
#
?
Sep 1, 2018 15:13
|
|
|
CPColin posted:The Devs should never see or care about the "Business Value" field
|
|
#
?
Sep 1, 2018 15:42
|
|
|
Vulture Culture posted:wait what "Business Value" is a concept that only the PO and the stakeholders should care about, while prioritizing the backlog. By the time the Devs see the backlog, the priorities should be set and nobody should be talking about Business Value any more.
|
|
#
?
Sep 1, 2018 16:16
|
|
|
Murrah posted:I wanted to post a really long post but thought against it. i do a lot of projects using new libraries etc at work to keep myself up to date, even if the project doesn't exactly call for it. you might not be able to get away with it for every project, but it definitely makes the 9-5 go by way faster.
|
|
#
?
Sep 1, 2018 16:21
|
|
|
The other complaint I’ve heard about just doing items in a queue is that business owners want to have a date when something will be done, so they can plan rollout and marketing around it. Saying that it will be done when it’s done doesn’t get much love.
|
|
#
?
Sep 1, 2018 16:39
|
|
|
smackfu posted:The other complaint I’ve heard about just doing items in a queue is that business owners want to have a date when something will be done, so they can plan rollout and marketing around it. Saying that it will be done when it’s done doesn’t get much love. it's rarely the coding that is the problem for me. like, in isolation, if i'm given a project with clear enough requirements that i don't have to keep asking for more information, and clear enough architectural direction, i can generally predict completion times to within a couple of weeks. what generally screws up my estimates are: a) incomplete requirements or knowledge of requirements b) having multiple concurrent projects and being inappropriately context shifted c) unforeseen/foreseen technical/architectural issue "when it's done" for me is a slightly better way of saying "i have no idea." I will always have an issue raised and an explanation for why I can't give an estimate, but generally if a situation is completely hosed, I won't even pretend to give estimates. this is to head off people asking every single hour/day when the problem will be fixed. i am perfectly happy to give estimates on projects with solid requirements, acceptance criteria, and technical direction, in an environment without heavy context switching.
|
|
#
?
Sep 1, 2018 17:08
|
|
|
If it's in progress, soon. If it's in todo, eventually. If it's in the backlog, never.
|
|
#
?
Sep 1, 2018 17:38
|
|
|
Murrah posted:I wanted to post a really long post but thought against it. Only the first part sounds like imposter syndrome. If you don't have a way to grow and learn, you will regress. Either try and get a mentor at your current job, start looking for a new one, or both.
|
|
#
?
Sep 1, 2018 17:48
|
|
|
Ok. Ill see if I can express myself clearly, because I'd like some kind of external feedback, even if its "thats the job/get used to it" or whatever. Sorry long post. I was asked to internationalize parts of the app at work to support the dd/mm/yyyy format for our customers in places in like UK/Australia (I mean the rest of the world other than the USA really) and it has been a terrible experience. The app is a PHP Symfony app thats been in production for about 8 years, most of that time as in house software tool and then the last 2-3 years as a public facing product, and has accumulated a large amount of code that relied explicitly/implicitly on mm/dd/yyyy dates. There is consistent/good practice in saving dates to the database in ISO format but the whole app is about tracking workflow and dates and there was just lots of little places where the american format was assumed. Large chunks of the app were able to be just "switched on" to the users culture to display the right format like tables generated within PHP (even this older version of Symfony had a fair bit of il8n support) but a significant amount of the app was not, especially anything interactive/where dates were submitted/processed/validated. To me the work was infuriating, trying to develop a solution that seemed 'elegant enough' while dealing with the quirks of things like PHP's strtotime handling dates as different formats depending on whether they are separated by / or -. I had to do similar things like processing everything to y-m-d but in many different context's places in the app. There are no automated tests or anything like that and I just instinctively felt like I was at risk of introducing several bugs for every change I was making and it felt like this for weeks and I would stew on this working remotely. It was basically assumed that I would find every instance in the app where dates would need to be coded differently through manual testing. Maybe I have just been having a bad few days but this was really demoralizing and I felt like programming was not fun anymore/wondering what I have got myself into. I would check in with my senior dev and frankly I was looking for some kind of assurance I was on the right track to finishing the project. Anyways the senior is a really smart but frankly quite serious introverted guy and would basically just say "I'd have to get all the context and look at it to say if this was a good strategy". The project kept extending and extending to about 4 weeks because I admit a failure that I simply had no idea how much code would have to be changed/ how many things broke when switching things like datepickers over to dd/mm/yyyy format. But then again I was just sort of let loose on the ticket without much thought into it by the PM/CEO I just last week was able to get it on to QA i.e all the known bugs were addressed. Im just left with the sinking feeling that there will certainly be more bugs and some that won't get discovered until production. I'm also wondering how this will be interpreted possibly 'against me' in assigning future tasks but meh Murrah fucked around with this message at 18:43 on Sep 1, 2018 |
|
#
?
Sep 1, 2018 18:33
|
|
|
Murrah posted:The app is a PHP Symfony app Found your problem  Murrah posted:There are no automated tests or anything like that Murrah posted:and I just instinctively felt like I was at risk of introducing several bugs for every change I was making Murrah posted:Im just left with the sinking feeling that there will certainly be more bugs and some that won't get discovered until production. You can pretty much bet on it. And as to whether it will be held against you, that depends on the culture at your company... I wanna say no, because they should understand that PHP app without any test is a crazy minefield, but then, they weren't smart enough not to have a PHP app without tests 
|
|
#
?
Sep 1, 2018 20:08
|
|
|
That sounds like a problem with scoping rather than a problem with your coding skills. Which is still an important skill to have as a software engineer, just not one that you necessarily acquire if you are purely writing code every day. It sounds like the estimate was based only on the desired end result, so you were guessing how many areas of the code you would have to work in. There should have been a research stage before that to identify as many of those places as possible so that the estimate could be more accurate. If that didn't happen then you were guessing and it's obviously pointless to criticise someone for guessing incorrectly. I'm not going to say "that's the job" because it's not universally true, but it is usually true if you work with a legacy code base that implemented a lot of bad practices, and management doesn't allow for that when planning changes. Which is going to be a lot of jobs.
|
|
#
?
Sep 1, 2018 20:19
|
|
|
CPColin posted:"Business Value" is a concept that only the PO and the stakeholders should care about, while prioritizing the backlog. By the time the Devs see the backlog, the priorities should be set and nobody should be talking about Business Value any more. In another sense, the entire value of user stories is to spell out business value and empower devs to ask the right questions. Vulture Culture fucked around with this message at 21:28 on Sep 1, 2018 |
|
#
?
Sep 1, 2018 21:20
|
|
|
Vulture Culture posted:Keeping the devs completely divorced from the intent of what they're building is pretty much a foolproof recipe for poo poo software (or worse, good software that never gets adopted successfully). They don't need to support every color of bikeshed that was suggested during the discussion process that led to that decision, but they should know what kind of results their work is expected to deliver to the organization so they can understand if they built the right thing or not. This is the whole reason so many product orgs are moving towards an OKR model—goals, metrics, and incentives are completely misaligned between layers of the business. Yeah, if I don't know why the business wants me to build something, I'll build exactly what they ask me to build. Sounds like a good thing except what they say they want is generally not what they actually want or need. It's not minor either. It can be on the level of, "We need you to build a machine that will shove people headfirst through this plate glass window." "...why do you want people shoved headfirst into a plate glass window?" "We need to get people into our store so they'll buy things!" "How about I build a door."
|
|
#
?
Sep 2, 2018 10:25
|
|
|
Che Delilas posted:It can be on the level of, "We need you to build a machine that will shove people headfirst through this plate glass window." "...why do you want people shoved headfirst into a plate glass window?" "We need to get people into our store so they'll buy things!" "How about I build a door." Hah, that is a perfect example and I agree that separating ~business value~ or the initial intent from developers is a recipe for disaster. Here is a (overly simplified) thing that used to happen when our api team was adjusting to working in sprints/with a backlog: Front-end: Backlog add: "Return disabled users from GET /users" Back-end: "Strange request, but ok, delivered" Front-end: "Disabled users still can't login" Back-end: "No of course not, but they are returned from /users like you requested." Front-end: "We don't care about that, we just wanted them to be able to log in and thought that would make them" Back-end: "The authentication token/login system has nothing to do with what is returned from that endpoint." ....... The moral of the story isn't that all front-end developers are idiots, but that we probably should have been talking with them directly to find out what they actually wanted to achieve. We have now sort of cheated and have a technical PO who used to be a BA, so my example wouldn't get past first draft anymore. We still have a general rule that if something "looks suspicious", we should find out what the requester ~really wanted~.
|
|
#
?
Sep 2, 2018 11:26
|
|
|
I was talking about shielding "Business Value" from the devs as shielding "the criteria the stakeholders and PO used to order the backlog" from the devs. I absolutely support capturing context and intent in backlog items. A properly empowered team would be allowed to reject a backlog item that felt too out-of-context, as above.
|
|
#
?
Sep 2, 2018 15:15
|
|
|
Che Delilas posted:Yeah, if I don't know why the business wants me to build something, I'll build exactly what they ask me to build. Sounds like a good thing except what they say they want is generally not what they actually want or need. god drat i have had this conversation so many times that it's painful.
|
|
#
?
Sep 2, 2018 15:40
|
|
|
Che Delilas posted:Yeah, if I don't know why the business wants me to build something, I'll build exactly what they ask me to build. Sounds like a good thing except what they say they want is generally not what they actually want or need. Asking these kinds of questions instantly puts you into the top echelon of devs at most places too.
|
|
#
?
Sep 2, 2018 17:08
|
|
|
CPColin posted:I was talking about shielding "Business Value" from the devs as shielding "the criteria the stakeholders and PO used to order the backlog" from the devs. I absolutely support capturing context and intent in backlog items. A properly empowered team would be allowed to reject a backlog item that felt too out-of-context, as above. Oh, that makes sense, and I agree.. I don't really care if Project A will bring us 1mil in 6 months but project B will bring us 4mil in 18 months. What I care about is that the product leaders have their poo poo together. Somewhat related: This week one of the managers decided he'd put together a meeting that included himself, me (the senior backend dev here), the lead front-end dev, the PO, every project manager, and the CEO. He wanted to talk about several completely unrelated bugs that have apparently been hanging around for a while. I think he called it a brainstorming session. I declined it and encouraged the other dev to do the same; I know most of the other managers including the CEO had other obligations. I understand that the organizer was unwilling or unable to articulate a concrete purpose of the meeting. So basically, dude didn't think his particular issues were getting enough attention and decided to waste some dev time by going around the process that's supposed to protect us from poo poo like this. I knew this was the case as soon as I saw the content of his meeting invite, which is why I declined it. This is what grooming is for, this is what the PO is supposed to decide - if you don't think your issues are getting enough attention, make your case to the rest of the management team and the PO specifically. Don't decide on your own that you get to ignore the only way we have to manage our completely maxed-out bandwidth. Fortunately we do have the authority to say no when someone tries this kind of garbage. We handle emergencies appropriately, of course, which is probably one reason why they trust us to make this kind of call. Doesn't stop the annoyance or the stress that comes from needing to maintain constant vigilance.
|
|
#
?
Sep 2, 2018 22:26
|
|
|
So I want a Cloud Platform for my NodeJS/MEAN apps but here's the thing: I'm familiar with Azure, but poo poo's expensive. I know Linux based solutions tend to be cheaper, but I'm absolutely useless with any kind of Linux-based CLI/SSH or anything like that. Maybe if I can remote into a GUI environment I might be safer, but that's a lot to ask for from a cheap Cloud provider. Basically I just want to be able to deploy Node apps, either Angular or React, and ideally Express/MongoDB via some sort of control panel or something. Any ideas? Ape Fist fucked around with this message at 12:34 on Sep 3, 2018 |
|
#
?
Sep 3, 2018 12:29
|
|
|
Ape Fist posted:So I want a Cloud Platform for my NodeJS/MEAN apps but here's the thing: I'm familiar with Azure, but poo poo's expensive. Have you looked into Azure Web Apps? They're pretty cheap.
|
|
#
?
Sep 3, 2018 13:39
|
|
|
Ape Fist posted:So I want a Cloud Platform for my NodeJS/MEAN apps but here's the thing: I'm familiar with Azure, but poo poo's expensive. Heroku like? The big 3 are all designed for someone who knows the ecosystem inside and out to deploy to and are complicated/expensive. They also all have some sort of ripoff of heroku in the form of AWS beanstalk, google app engine, and azure web apps, but have weird design limitations usually. I’d steer clear of AWS, as it’s a bit intimidating and confusing to get started without systems experience. Google has some nifty features like cloud console that make it more dev friendly to start with, but it still might be overkill. My experience with Azure is that it’s confusing and weird but that might just be me getting used to it’s vocabulary. But in the end the cheapest option is probably doing what your forefathers have done and getting a DigitalOcean droplet and following random guides on the internet to install what you need on Ubuntu.
|
|
#
?
Sep 3, 2018 13:53
|
|
|
I'm actually pretty handy with Azure/VSTS from using it at work. I'm fully capable of deploying a working MEAN application to Azure but ughhhhhhhhhhhhhhhh Cosmos DB is incredibly expensive. I actually love Azure, just not wild about my MEAN app nobody is going to use costing like £90 a month because I have to assign a minimum amount of request units to cosmos and it's just sitting there burning money forever. But I guess if I'm just using it to host a backendless front end app it's probably going to be pretty cheap.
|
|
#
?
Sep 3, 2018 14:44
|
|
|
why is cosmos db so stupid expensive? To drive SQL sales?
|
|
#
?
Sep 3, 2018 14:58
|
|
|
freeasinbeer posted:
Less so than AWS? I was helping a friend out with some AWS stuff recently, without having much experience with it, and I felt like I needed a "AWS moon-speak to normal terminology" translation service. It felt oddly cultish and was really irritating. I need to host a PaaS website. Azure: Azure Web App. AWS: Elastic Beanstalk. What the gently caress? Why? I need to hook a custom domain up to my Elastic Beanstalk site. Azure: "Custom Domain" section of my web app AWS: Route 53. What the gently caress? Why?
|
|
#
?
Sep 3, 2018 15:00
|
|
|
I found Azure pretty straightforward compared to AWS and it's got the same sort of user interaction logic most Microsoft products have. I found a similar problem with Digital Ocean in that I didn't know what any of their made up terms for poo poo actually meant and I was like welp, guess I'm not using this loving thing.
Ape Fist fucked around with this message at 15:55 on Sep 3, 2018 |
|
#
?
Sep 3, 2018 15:07
|
|
|
New Yorp New Yorp posted:Less so than AWS? I was helping a friend out with some AWS stuff recently, without having much experience with it, and I felt like I needed a "AWS moon-speak to normal terminology" translation service. It felt oddly cultish and was really irritating. see also: the AWS Icon Quiz
|
|
#
?
Sep 3, 2018 15:52
|
|
|
That's appalling. Why even bother having iconography? It's worse than death metal band logos.
|
|
#
?
Sep 3, 2018 16:16
|
|
|
New Yorp New Yorp posted:Less so than AWS? I was helping a friend out with some AWS stuff recently, without having much experience with it, and I felt like I needed a "AWS moon-speak to normal terminology" translation service. It felt oddly cultish and was really irritating. I guess my issue with Azure is that they have 3 different versions of everything and I don’t know which one I am supposed to use. Pricing also feels impossible to figure out. I already use two clouds in day to day work, so I don’t really have a ton of desire to figure out Azure on top. Also rightly or wrongly I filter out Microsoft products because I use Linux day to day, and Azure heavy shops seem to typically be Windows heavy shops with BOFHs/legacy mindsets that are a pain to interact with. Now that isn’t to say all Azure/Windows shops are bad, but the ones I interact with or have exposure to are terrible. As far as AWS, I’ve been using it for almost a decade so I am just used to it now.
|
|
#
?
Sep 3, 2018 20:22
|
|
|
freeasinbeer posted:Pricing also feels impossible to figure out. This is probably my biggest complaint with Azure, tbh. You get a bunch of free credit, which you will immediately piss through in a month because the default configuration assumes you want enough resources to run a medium-scale business, not a learning experiment.
|
|
#
?
Sep 3, 2018 21:56
|
|
|
New Yorp New Yorp posted:Less so than AWS? I was helping a friend out with some AWS stuff recently, without having much experience with it, and I felt like I needed a "AWS moon-speak to normal terminology" translation service. It felt oddly cultish and was really irritating. There's AWS in Plain English.
|
|
#
?
Sep 3, 2018 22:02
|
|
|
Boiled Water posted:why is cosmos db so stupid expensive? To drive SQL sales? Isn't it supposed to be a globally distributed DB with SQL support? There's some blog with a developer interning at Microsoft for a year or so and masterbating furiously over how wonderful it is and so much better than Google's Spanner.
|
|
#
?
Sep 4, 2018 01:37
|
|
|
MrMoo posted:Isn't it supposed to be a globally distributed DB with SQL support? There's some blog with a developer interning at Microsoft for a year or so and masterbating furiously over how wonderful it is and so much better than Google's Spanner. (just the fact that it's a blob db will tell you it's not better than spanner) (no joins allowed, no groupbys allowed)
|
|
#
?
Sep 4, 2018 02:22
|
|
|
Ape Fist posted:I'm actually pretty handy with Azure/VSTS from using it at work. I'm fully capable of deploying a working MEAN application to Azure but ughhhhhhhhhhhhhhhh Cosmos DB is incredibly expensive. Are you dead-set on using Cosmos? It is expensive but there are other options for hosted DBs on Azure. The "basic" tier Azure SQL instance is like $5/month, and if you need a document store you could use something like this to get Mongo running on Azure.
|
|
#
?
Sep 4, 2018 13:55
|
|
|
freeasinbeer posted:I guess my issue with Azure is that they have 3 different versions of everything and I don’t know which one I am supposed to use. Pricing also feels impossible to figure out. I already use two clouds in day to day work, so I don’t really have a ton of desire to figure out Azure on top. Also rightly or wrongly I filter out Microsoft products because I use Linux day to day, and Azure heavy shops seem to typically be Windows heavy shops with BOFHs/legacy mindsets that are a pain to interact with. My main complaint with Azure was that most of the hosted services are reasonable to manage, but everything around VMs (particularly Linux VMs) was totally half-baked at launch. There was no metadata API until less than 18 months ago, meaning all your host-specific VM bootstrap stuff had to be handled on another channel using something like Terraform provisioners. The storage offerings don't make any sense to me, and I still don't understand when I need to use what kind of storage for what kind of disks on what kind of VM to not have it spit errors at me. The first time I ported our standard AWS/OpenStack image using Packer, I spent days trying to get it to work, because the VM will never report as online unless the following two conditions are met: a) their proprietary agent starts up and b) the VM writes output to the serial console. I get it, sort of, but some sort of something explaining the problem would have been terrific. (Microsoft's company-wide, cultural penchant for awful error messages has not improved on Azure.) Scripting outside of PowerShell is still kind of user-hostile because a number of operations in the console are synchronous, and there's no option to just not wait on a result. As an example: VM reboots will not exit until the VM is back online and it has a success status to report on the VM bringing itself back up, so good luck to you if you need to do thousands of them concurrently. It's gotten better, but Amazon and Google are in a price war that Microsoft has been sitting out, and you'd have to have some pretty specific requirements for adopting Azure on VM-heavy workloads to be a good business decision. Vulture Culture fucked around with this message at 14:07 on Sep 4, 2018 |
|
#
?
Sep 4, 2018 14:02
|
|
|
freeasinbeer posted:I guess my issue with Azure is that they have 3 different versions of everything and I don’t know which one I am supposed to use. This applies equally to AWS, FYI. You're probably just used to it because you've been using it long enough, but we just started using it recently and it's not that straightforward. I'm thinking specifically about a ten minute discussion we had about which storage option was right for a new workload I was going to saddle a service with. Much of that was synchronizing knowledge about what the various options are supposed to be for.
|
|
#
?
Sep 4, 2018 17:36
|
|
|
Ape Fist posted:This is probably my biggest complaint with Azure, tbh. for SQL Server it's just insane. DTUs, vCores and prices so high it's actually worrying.
|
|
#
?
Sep 4, 2018 18:33
|
|
|
for the top of the line sql server paying retail, assuming a pizza box server, it costs more to license the thing w/ the top of the line license than to send it to space it's common to get 90, 95, 99% discounts tho
|
|
#
?
Sep 4, 2018 19:21
|
|
|
|
| # ? May 23, 2024 12:19 |
|
|
Azure has Web app + SQL that is not terrible.
|
|
#
?
Sep 4, 2018 19:53
|
|
