4 ms·

> We are aware that "it doesn't run". ...what? You have an application that literally doesn't run and you've been working on it for months?

mewpmewp2 2y ago
Presumably there is some legacy integration or blocker why it doesn't run.
- Sponge5 2y ago
  Even so, working as an engineer on a project that doesn't run as a whole for more than a week, let alone months.. It's very understandable that management is nervous. Cut things away until you have something that runs is step one...
bruce511 2y ago
It's probably an over-simplification to say "it doesn't run". Firstly, this is a conversion of existing code to a later version of the compiler. (a 25-year later version). So it's not "writing a new app" - it's making it all compile (and run) under the new compiler. To be clear - it does compile. That was a pretty small part of it. And it does "run" in the sense that the Exe starts, gets through authentication and so on. All the Utilities run, and the 2nd-largest system is running. The primary program consists of around 15 threads all running simultaneously. The code was originally written using a co-operative threading model, and it's now going to use a preemptive model. The program that "doesn't run" is effectively (not yet) starting all the threads correctly, in the right order, and at the right pace. (There is a lot of cross-thread communication and cross-thread UI interactions.) So yeah, it's a giant big ball of smoosh. This is not a project for the faint-hearted. But we're actually nearing the end of the tunnel now - every day progress is being made. But until we can show the "whole program" running, to management progress seems to be binary. (It either works or it doesn't.)
- rrr_oh_man 2y ago
  > But until we can show the "whole program" running, to management progress seems to be binary. (It either works or it doesn't.) That's insufficient communication from delivery (= you), imho. When a tunnel gets dug, you also don't always know what you're going to encounter (surprise water table? rock? skeletons?), but you make a plan to the best of your current knowledge, track your progress against it, and adjust where needed. If you cannot do that, it looks like you're just trying to power through the project blind based on gut feeling alone.
  - skydhash 2y ago
    Same as if you're a contractor. You propose a plan so they can track progress, estimate it so they plan for dependent stuff, and adjust the plan/timetable when issues arise (and communicate plenty and plainly). While I don't know GP's project context, I'd go with a test checklist and measure against that.

mewpmewp2 2y ago

Presumably there is some legacy integration or blocker why it doesn't run.

Sponge5 2y ago
Even so, working as an engineer on a project that doesn't run as a whole for more than a week, let alone months.. It's very understandable that management is nervous. Cut things away until you have something that runs is step one...

bruce511 2y ago

It's probably an over-simplification to say "it doesn't run". Firstly, this is a conversion of existing code to a later version of the compiler. (a 25-year later version). So it's not "writing a new app" - it's making it all compile (and run) under the new compiler. To be clear - it does compile. That was a pretty small part of it. And it does "run" in the sense that the Exe starts, gets through authentication and so on. All the Utilities run, and the 2nd-largest system is running. The primary program consists of around 15 threads all running simultaneously. The code was originally written using a co-operative threading model, and it's now going to use a preemptive model. The program that "doesn't run" is effectively (not yet) starting all the threads correctly, in the right order, and at the right pace. (There is a lot of cross-thread communication and cross-thread UI interactions.) So yeah, it's a giant big ball of smoosh. This is not a project for the faint-hearted. But we're actually nearing the end of the tunnel now - every day progress is being made. But until we can show the "whole program" running, to management progress seems to be binary. (It either works or it doesn't.)

rrr_oh_man 2y ago
> But until we can show the "whole program" running, to management progress seems to be binary. (It either works or it doesn't.) That's insufficient communication from delivery (= you), imho. When a tunnel gets dug, you also don't always know what you're going to encounter (surprise water table? rock? skeletons?), but you make a plan to the best of your current knowledge, track your progress against it, and adjust where needed. If you cannot do that, it looks like you're just trying to power through the project blind based on gut feeling alone.
- skydhash 2y ago
  Same as if you're a contractor. You propose a plan so they can track progress, estimate it so they plan for dependent stuff, and adjust the plan/timetable when issues arise (and communicate plenty and plainly). While I don't know GP's project context, I'd go with a test checklist and measure against that.