estimation   3925

« earlier    

The Math behind Project Scheduling, Bug Tracking, and Triage | USENIX
Many projects have poorly defined (and often overridden) priorities, hopelessly optimistic schedules, and overflowing bug trackers that are occasionally purged out of frustration in a mysterious process called "bug bankruptcy." But a few projects seem to get everything right. What's the difference? Avery collected the best advice from the best-running teams at Google, then tried to break down why that advice works—using math, psychology, an ad-hoc engineer simulator (SimSWE), and pages torn out of Agile Project Management textbooks.

We'll answer questions like:

• Why are my estimates always too optimistic, no matter how pessimistic I make them?
• How many engineers have to come to the project planning meetings?
• Why do people work on tasks that aren't on the schedule?
• What do I do when new bugs are filed faster than I can fix them?
• Should I make one release with two features or two releases with one new feature each?
• If my bug tracker is already a hopeless mess, how can I clean it up without going crazy or declaring bankruptcy?
project_management  software  estimation  agile  kanban  bug_tracking  bestpractices  is:video 
2 days ago by cdzombak
Estimating project costs? If statements should cost $10,000 each • Dave Rupert
Dave Rupert:
<p>“Whoa, whoa, whoa, that would never work” I hear you say. But there’s never been an easier way to convey the scope and cost of a project than if-statement based billing. What is an if-statement? An if-statement is the most essential unit of business logic. A small piece of logic that will linger in your codebase for the life of the entire project. Larger software applications have more business logic, thus are more expensive. We can use if-statements as a proxy for complexity and bill accordingly. At the end of the day developers can count up the number of if-statements and invoice the corresponding cost centers.

What about small projects, you say? Well, the beauty of this is something simple like a blog is actually free! Free website? Yes, please.

But let’s say your app has a logged-in or logged-out state, well, that’s at least 2 if-statements. Starting price: $20,000. Never before has it been this easy to price and scope out complex stateful apps!

Do you build Component Systems? Simple static components are free. But most components increase their cost due to the The Nine States of Design. Each component likely has a mix of “none”, “one”, “some”, “too many”, “error”, and “done” states. That’s a lot of logic and use cases packed into a little module, so it’s gonna cost ya. But you’ll rest assured that you’ve covered all your bases as well as billed appropriately.

Need an if-statement with 2 conditionals? Look, I’m not a scam artist so I’ll give you the second conditional at half-price. But if it gets any more complex than that and we <a href=">have to build a big juicy Karnaugh Map</a>, that gets into bitwise operators (which are generally a terrible idea in JavaScript) and will double the cost per switch case.</p>

This is both hilarious and yet also true.
Software  programming  estimation  project 
4 days ago by charlesarthur
If Statements Should Cost $10,000 Each
An exciting new thought technology for software project estimation
fridayfrontend  javascript  conditionals  estimation  humor 
16 days ago by spaceninja

« earlier    

related tags

(popular  2018  3d  95  agent-based_models  agile  alancooper  algorithm  analysis  approximation  arc  asymptotics  aurore  bestpractices  bias  biomass  blogpost  bookmarks)  bug_tracking  business  camera  cards  career  charts  communication  complexity  conditionals  confidence  controlsystems  controltheory  cost  counting  crime  cto  culture  data-science  data-structures  datascience  dataset  deep-learning  delaigle  density-estimation  density  development  devops  divergence  divergences  economics  entropy  error  essentials  estimates  evolution  facebook  fermi  filter  filters  flow  forecasting  france  fridayfrontend  github  golang  have_read  high-dimensional_statistics  humor  immobilier  in_nb  indirect_inference  inference_to_latent_objects  informationtheory  interesting  interpretation  is:tweet  is:video  javascript  job  kalman  kanban  kernel  kullback-leibler  lang:de  latency  leadership  library  likelihood  machinelearning  magic  management  markov_process  math  matlab  maximum  methodology  ml  moderation  montecarlo  network_data_analysis  neuf  non-equilibrium  nonparametric  observation  on-line  order-of-magnitude  pandas  parameter  perception  philosophy  planning  planung  pm  pnas  points  pose-estimation  pose  predicition  prediction  prioritisation  prix  probabilistic-counter  probabilistic-model  probability  probabiltiy  process  product  productivitiy  productivity  productmanagement  progerammer  programming  project-management  project  project_management  projectmanagement  province  psychology  python  quote  rare-event_simulation  re:hyperbolic_networks  retro  robins.james  science  scrum  serverless  simulation  sizing  sketch  software-development  software-management  software  software_engineering  softwareengineering  standup  statistical_inference_for_stochastic_processes  statistical_mechanics  statistics  story-points  story  tdd  team-tools  team  teams  tech  technical-debt  technicaldebt  thoughts  time_series  tips  to:nb  to_read  tools  tuning  tutorial  ui_ux  uncertainty  vente  visualisation  visualization  wisdom  work  xp 

Copy this bookmark: