Hacker News Reader

Comment by jl6 | Hacker News Reader

It’s fun to point and laugh, but is this really just the consequence of an unprecedented health emergency demanding a rush job?

As far as bugs go, it doesn’t sound that bad. They didn’t lose data - they just processed it late? And they spotted it within days/weeks, and have a workaround/correction already? And it’s only the reporting that was wrong, not the more important part where they inform people of results?

I’d rather have this system now than be waiting for the requirements analysis to conclude on the perfect system.

View on HN

There's a test and trace system that is supposed to phone everyone with a positive test result and get details of all their close contacts (people they sat in an office with, restaurants they visited, etc), who are in turn contacted.

A delay in the test results means the contact tracing could not happen, and people will have been going around spreading the virus they caught off the original, known cases.

Also, the whole test and trace system has been in the news a lot recently here for various failings, and things like this will just further knock people's confidence in it.

> A delay in the test results means the contact tracing could not happen, and people will have been going around spreading the virus they caught off the original, known cases.

A recent piece in The Atlantic argues that we got it all wrong, we should test backwards (who infected the current patient) instead of forwards.

This particularity is caused by the dispersion factor K which is a metric hidden (conflated) by the R0, meaning that the same R0 would be handled differently depending on the K factor.

A related suggestion was that we should focus more on super-spreaders as COVID doesn't spread uniformly. That's why the probability of finding a cluster is higher looking back than forward - most people don't get to actually spread the virus.

https://www.theatlantic.com/health/archive/2020/09/k-overloo...

You’re right - but the biggest issue is clearly one of experience. Someone was put into a critical role they weren’t prepared for - a position in which their failures cost lives. They don’t have a redundant process, or error checking, or even someone familiar with spreadsheets.

That’s the scandal: it’s the type of basic error that it screams “they’re not handling this well.” You wouldn’t tolerate your new SWE colleague asking “What’s an array again?”

If publishing the whole set of data was delayed by excel crashing - fine. But silent failure because they ran out of columns? Cmon...

I would say it's error from a rushed (and understandably so) job, I've been involved in a number of Covid related projects at the US county and state level. The teams and departments are having to move very fast with difficult technical problems. Keep in mind, it's not like they were just sitting around with nothing to do pre-covid. Basically, when the pandemic hit, everyone got a second job to go along with their day job.

I think that's too generous. It would indeed have been excusable for perhaps the first 2 weeks. But at this point, they've had an effectively unlimited budget, likely access to engineers working for big consultancies and big tech who would have been quite happy to literally drop everything to help out and 6 MONTHS to build a system that worked. And they are still using an Excel spreadsheet.

Well, I don't see why developing an application for transforming csv files to xls files would go faster than developing an application that uploaded csv files to a database.

I could see that i could be a rush job problem, but in this case they're not gaining any time.

Good points

Partly I think the Excel bit is news because Excel (and mistakes made with Excel) is easily relatable to people. But bugs always come up and most other bugs are just as stupid. If it had been an off-by-one loop error in some C code somewhere it would be just as dumb but you'd get none of the facepalm memes all over Twitter.

Off by one errors are all alike; every excel error is an error in its own way.

Not when it mixes up US and UK date formats when you import a CSV.

Not when it turns phone numbers into integers and strips off the leading zero.

This is the root of the problem isn't it? The wrong tool for the job.

Tolstoy keeps on giving ...

They had a budget of GBP 36 million. It's pretty bad.

Serco received 45.8 Million in March (not disclosed until April), and I'm pretty sure they've had more since.

What’s your source for that budget? The error was by PHE, not Serco.

You need to back that number up with real facts, as this article is regarding the PHE handling of reporting.

This is what it most likely is. Analogous to starting out on an easy to hack together tech stack, succeeding, and dealing with tech debt when something goes wrong.

I guess its easier to understand since everyone uses excel, however it does end up giving a halo of blame to excel, as opposed to human processes.

No. There is difference between a little rush job, and an incompetent rush job. I’m only speaking for my own experience, but when put under pressure most competent teams don’t end up throwing away such important data. I’ve been incompetent for most of my career, but it’s given the serious of task this is inexcusable. I don’t know how things work elsewhere, but checking that the same number of unique records(by some composite key) are ingested as exported isn’t an advanced concept. Know the value of the data you’re handling. Marketing data <> health data. If health data doesn’t line up: you figure out why and solve it. If you don’t have anyone available to do so then the problem is managerial/structural.

That has been the COVID experience in Britain in the last couple of months: the government is spending a ton of money in training new people and creating new processes, which is causing a ton of slowdowns and bugs.

In the grand scheme of things, having problems because of inexperienced people is much better than not having enough people like during April and May.

Reply on news.ycombinator.com