Hacker Newsnew | past | comments | ask | show | jobs | submit | setopt's commentslogin

It was good on a long time scale, but I think the parent poster refers to the short term. If I recall correctly, during the early Industrial Revolution the average life span decreased, child mortality went through the roof, and malnutrition meant adults lost their teeth in their early 20s at best. That was… worse. It took time for the revolution to become a net-positive for the average person (which I certainly wouldn’t dispute).

In my experience, it’s hard enough to make students check their school email in the first place. Let alone filter it.

As a ugrad, and later a PhD student teaching, everything is explained the first day. If you can figure it out you just fail the class (or go to office hrs to get help, etc).

As an associate professor, I do explain things the first day, but I am certainly not permitted to fail students as a consequence of not checking their email daily.

Even if they didn’t hand in an assignment at all, without any reason provided, I’m required by regulation to offer them a second chance to pass that assignment.

The students’ rights are quite strong here (Northern Europe), which I generally support, but it has some downsides.


Interesting. I remember very strict rules on turning in programming assignments (as a student, and later TA). On time, printed properly, in a specific envelope, labeled as specified in the right location.

Just to add one more data point, we also use Canvas at my university. The deadline for submitting who are eligible (i.e. passed compulsory assignments and labs) to take the exam was yesterday, and I couldn’t meet that deadline because Canvas went down. I usually do corrections offline so I have backups of my own evaluations, but these are courses with many teachers and many TAs, so Canvas is the way we sync our assessments.

I guess what surprises me the most is that it’s even legal for schools to outsource the core of what they do to some random tech company.

Either way, they were under no obligation to adopt this garbage technology regardless of whether it’s available, so this is 110% on them.


I’m sorry… is your view here that you can’t believe it is legal for a school to purchase software or pay someone to host software for them?

You are aware that you are posting on Hacker News, a forum for people who make their living selling software and the expertise to host it?


The alternative would be that each school develop their own platform for this, which also isn't very good use of their time and money?

Edit: No idea why this was down voted so much. I'm not defending Canvas, just wondering what the alternative would be.


> The alternative would be that each school develop their own platform for this

I worked at a university which did exactly this, in the UK.

It was a bespoke platform which integrated incredibly well with the rest of the systems the university used because it was designed from the ground-up to meet the institution's needs, there were regular user groups involving academics to understand what features needed to be built/worked on etc. At one point it was all OSS on GitHub too, in case other universities could've found it useful. It handled plagiarism detection (integrating with Turnitin), marking, exam grids, coursework submissions and feedback, seminar allocations, personalised timetables & mitigating circumstances.

The in-house dev team was vastly cheaper than anything SaaS would've cost, as well. It also maintained software for on-campus parcel deliveries, online exams, opinion surveys, a mobile app for students/staff, the SSO system, the course catalogue, car parking permits, a content management system and more.


That sounds like a dream.

My (also UK-based) university has been working on a new student records management project for years that's been incredibly ill-fated. It's destined to replace all their current systems and the first module module was meant to launch last year, except it thoroughly failed testing and nobody has heard anything about it since.

No idea how long it'll take to pull through. I don't believe it's an in-house effort.


In-house bespoke software sounds reasonable, and multi-customer SaaS sounds reasonable, but outsourced bespoke software sounds like a complete dumpster fire:

End users who report problems:

* are ok with IT level 1 telling them IT level 3 is working on it with velocity appropriate to keep their jobs,

* are ok with IT level 1 telling them ${vendor_of_well-known_solution} is working on it with velocity appropriate for many customers, but

* are not ok with IT level 1 telling them ${vendor_of_bespoke_solutions} is working on it with velocity appropriate for one customer (if they even still exist).


This sounds like a great opportunity for students to gain hands on experience with real software engineering work as well.

They do not need to develop it, but host an existing software on their infrastructure maybe...

The alternative could be to self host.

https://github.com/instructure/canvas-lms/wiki/Production-St...

Or maybe consider not following the herd, and use a much simpler but sufficient system that can be self hosted, if available.


The alternative is FOSS.

Seems like instructure canvas is FOSS: https://github.com/instructure/canvas-lms/tree/master

If your line is GPL rather than AGPL there's Moodle.

But you do then have to have a sysadmin capable of managing an enterprise grade LAMP stack.


Canvas already is AGPL, though?

So it can be used by multiple universities who share the maintenance. That is my point: Not everybody has to develop their own.

Um. This is the forum for an industry that outsourced its entire core of what they do to Microsoft (GitHub).

I had to check it up too. Appears to be a synonym for "inverting" used in some fields like biology and medicine.

It makes economic sense because they require a large initial investment (CAPEX), but low cost per year to keep functioning for many decades (OPEX). In contrast to say wind or solar, which are smaller CAPEX but higher OPEX.

So when you compare average cost per year over the complete expected lifetime of the plants, nuclear is good, but when you compare the up-front cost to build it, yeah it looks bad.

Another thing is that nuclear in the US is far more costly than in e.g. France. The key is that France standardized a few reactor designs that they kept building again and again, which made both construction and maintenance cheaper over time. While in the US, each nuclear plant is a unicorn, which can perhaps result in better individual designs but ends up more expensive.


Unfortunately France can no longer build nuclear plants cheaply either. All of the recent nuclear plants built by the French state owned company EDF in France, Finland, and the UK have seen enormous cost and time overruns.

Cumulative emissions matter. We simply don’t have the time to wait the 20 years it takes to build new nuclear plants.


Source please? The numbers I have seen of real opex paint a different picture. In general, nuclear plants close because of cost.

Googling I see numbers like $11/MWh for onshore wind [1] and $5-$10/MWh for solar [2] while nuclear is around $25/MWh [3].

[1] https://emp.lbl.gov/publications/benchmarking-wind-power-ope...

[2] https://emp.lbl.gov/publications/benchmarking-utility-scale-...

[3] https://www.nei.org/getContentAsset/47fa8caa-9b0d-4029-932c-...


Also wow, for Solar the property taxes are OpEx. So if there's more sunlight because of good weather these "Operating expenses" decrease because they're based on taking fixed costs like property taxes and just dividing them by power output that's unrelated.

I assume property taxes for a gas turbine are likewise OpEx but they just disappear in the noise of buying enormous amounts of methane as fuel.


How long until we have to solve a captcha per message to counter that?

Are captchas still effective against modern LLMs?

For this use case it matters a lot less if LLMs can solve it. As long as it costs you more to solve the captcha than it costs your adversary to serve it to you, it is still (some what) effective.

They are if your goal is to burn their GPU time but instead of hundred requests a second you're busy solving captchas

Claude can burn tokens solving the captcha for me. Double the effect.

Since it’s an upgrade of markdown, you should have called it "markup".


This sounds to me like it might benefit from some sort of "hypertext" functionality to allow for easy linking of documents


Now you're cooking with gas. Maybe it could be some sort of semantic markup language so we can separate and annotate things like titles, headers, links, and all of that stuff.


I would like to include some dynamic content in my documents. Could we include some kind of simple scripting language?

Maybe call it something like that really popular, Java language? But of course, have it share no concepts with Java, because that would be too straightforward.


This is very interesting. I imagine the links of these documents to each other could be visualized as a web of some sort.


"updoc" is still my favourite joke name. A long time ago (predating E lang's updoc afaict) I wrote a toy markup for semi-technical docs, named so with the specific intention of dropping it casually into conversation. Still funny :D

Price accordingly.


I don’t disagree, I just moved back to Linux from macOS myself (Tahoe was the last drop for me).

But did you try Homebrew and its extensions? It works pretty well for managing both terminal and GUI apps, and has some useful extensions like Brewfile, MAS, etc. Its not perfect, but for single-user Macs with an up-to-date OS version, it works quite well.


> I would welcome a version of it that stripped out everything but that, and just replayed the last few pages of scrollback on reattach.

Like `dtach`?


Thanks. Dtach is definitely closer to what I want than tmux, I'll be trying it out this week.


I had a similar moment a few years ago. That Google Maps pop-up was what caused me to first switch to de-googled Android, and once that turned out to be a hassle after a couple of years, switch to an iPhone without Google stuff. (On Android, Google is a location provider, so blocking their access is much harder.)


>On Android, Google is a location provider, so blocking their access is much harder.

https://grapheneos.org/features#network-location

Their approach encompasses GNSS location, too. Nothing Google required.


True. Sidenote: they are still however push notifications provider, so good luck getting rid of them completely (unless you're fine with not getting the notifications). MicroG is awesome wrt. that as you can turn it on/off as you wish, and it just works. GrapheneOS however only supports Google services in sandbox, but the notifications work sporadically IME (maybe because I keep turning them off and on... not sure). So... Pick your poison.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: