Hacker News Reader

fcurtson London may have gone into a Covid-accelerated decline

"Maybe" at individual level becomes "certainly" at city, state, or nation level.

fcurtson London may have gone into a Covid-accelerated decline

I have (both my parents died from cancer in their 50s). How is this related to this topic? Lockdown or not, crowding and not wearing a mask is irresponsible - you aren't just putting yourself at risk, but also the people around you.

fcurtson London may have gone into a Covid-accelerated decline

Except for the ones that will end up fighting the worst illness in their lifetime. It may not be them but their parents or grandparents.

fcurtson Don't write documentation in Markdown

I had the same gripes with RST. Then I found Asciidoc(tor) and never looked back. Asciidoc is as easy to learn and use as Markdown and more powerful than RST. Best combined with Antora, Asciidoc's answer to Sphinx.

fcurtson Don't write documentation in Markdown

For us, Asciidoc(tor) is the answer. As easy to learn and use as Markdown, more powerful than RST. Best combined with Antora, Asciidoc's answer to Sphinx.

fcurtson Don't write documentation in Markdown

Asciidoc(tor) is far easier to learn, read and write than RST (I have extensive experience with both). Our developers love it. We use it together with Antora, Asciidoc's answer to Sphinx.

fcurtson Don't write documentation in Markdown

The Ruby implementation (Asciidoctor) is a separate implementation by a different team. I recommend to use it together with Antora (Asciidoc's answer to Sphinx).

fcurtson Covid-19: the harms of exaggerated information and non‐evidence‐based measures

I don't buy that. Judging from the actions of governments around the world, many of which made drastic course corrections based on recent findings, the consensus is that the risk of millions of deaths due to underreaction is far greater than the risk of millions of deaths due to overreaction. Of course, it will be important to continuously reassess the situation and make course corrections in either direction based on new findings. In the meantime, governments should do everything they can to mitigate financial hardship and keep the economy alive. Unfortunately, there is no way out of this that won't cause significant pain.

fcurtson Covid-19: the harms of exaggerated information and non‐evidence‐based measures

In that case, why wouldn't the same hold for underreaction?

fcurtson Covid-19: the harms of exaggerated information and non‐evidence‐based measures

Hospitals in Italy are currently reporting dramatic shortages of many kinds.

fcurtson Covid-19: the harms of exaggerated information and non‐evidence‐based measures

They buried or cremated them. Unsurprisingly, the system can't handle a sudden one-of-a-kind spike.

fcurtson Covid-19: the harms of exaggerated information and non‐evidence‐based measures

> An argument in favor of lockdowns is that postponing the epidemic wave (“flattening the curve”) gains time to develop vaccines and reduces strain on the health system. However, vaccines take many months (or years) to develop and test properly.

It's not only about vaccines but about treatments in general. There is a good chance that we'll have a drug that significantly improves outcomes long before we have a vaccine. I don't understand how a peer-reviewed paper can overlook this.

fcurtson Ask HN: What chance that we wouldn't notice Covid19 if we hadn't tested for it?

https://reason.com/2020/03/17/italian-daily-death-rate-up-20...

The evidence is overwhelming. Questioning it is an extraordinary claim. Extraordinary claims require extraordinary evidence.

I won't comment further.

fcurtson Ask HN: What chance that we wouldn't notice Covid19 if we hadn't tested for it?

Please stop spreading FUD. Thank you.

fcurtson Ask HN: What chance that we wouldn't notice Covid19 if we hadn't tested for it?

> If less than 30% of the additional deaths this year were attributed to corona virus and 30% lies within a reasonable variance of this number over the years, then we wouldn't have noticed, or not?

You seem to be misinterpreting the statement, which says that even though "only" 30% of additional deaths are proven to be caused by the virus, the real number is likely higher.

It should be obvious that the situation in northern Italy is far beyond "reasonable variance". Google is your friend.

fcurtson Ask HN: What chance that we wouldn't notice Covid19 if we hadn't tested for it?

> how likely is it, that we would have never found anything being wrong just based on the people that die from Pneumonia related issues this year if we had never found Covid19?

0%. Just read the reports from Wuhan and Italy. For example, citing from https://www.worldometers.info/coronavirus/country/italy/:

March 20 (GMT)

Italy: in the city of Bergamo, there were 108 more deaths in the first 15 days of March this year compared to 2019 (164 deaths in 2020 vs. 56 deaths in 2019) according to the mayor of the city Giorgio Gori. During this period, 31 deaths were attributed to the coronavirus (less than 30% of the additional deaths this year)

"There are significant numbers of people who have died but whose death hasn't been attributed to the coronavirus because they died at home or in a nursing home and so they weren't swabbed," said the mayor [source]

March 19 (GMT)

Italian army trucks form a long line in the center of Bergamo to transport the coffins from the Bergamo cemetery to cremation sites in other regions, as morgues can't cope with more coronavirus deaths

Obituaries in the local newspaper went from 1.5 pages (on Feb. 9) to 10 pages (on March 13)

fcurtson Ask HN: Why are there so few confirmed COVID-19 cases in San Francisco?

Some observations:

* Santa Clara County has over double the population of SF County.

* Santa Clara County has always been ahead of SF County in confirmed cases.

* According to https://projects.sfchronicle.com/2020/coronavirus-map/, the ratio has stayed roughly the same over the last two weeks (3:1).

* Interestingly, LA County had fewer confirmed cases than SF County until 11 days ago.

fcurtson California governor issues statewide 'stay at home' order

> Why do people think that waiting 2 weeks or 2 months will solve it?

Who said it will solve it? It will flatten the curve as it already has in other countries, hopefully avoiding a catastrophic overload of the healthcare system.

fcurtson Ask HN: Why don't homeless people from SF move to another city?

When you are broke, live on the street (which is a very tough life everywhere in the world), and possibly fight with mental illness or addiction, moving to a more affordable place (which will still be unaffordable for you, esp. in California) won't magically solve your problems. Without the right support, it will be difficult to find your way back to a normal life.

fcurtson Negotiations Failed: How Oracle Killed Java EE

> my guess is that I can't use graalvm to produce a binary which I can sell without either open sourcing everything I've made or paying too much to Oracle.

GraalVM CE is licensed under GPL 2 with Classpath exception (https://github.com/oracle/graal/blob/master/LICENSE), specifically so that binaries created with GraalVM (which contain the GraalVM runtime) don't fall under the GPL.

fcurtson Clojure 1.10 release

A more feasible option is to compile Clojure byte code to native code with GraalVM (http://www.graalvm.org/docs/reference-manual/aot-compilation...). GraalVM's native compiler still has some limitations but is coming along nicely.

fcurtson Grain: A strongly-typed functional programming language for the modern web

Dart 2 has a sound type system and catches this at compile time.

fcurtson Do I need to pay for Java now?

Apparently some links broke, but nothing was taken offline: http://www.oracle.com/technetwork/java/javase/downloads/jdk8...

> Can you source that claim?

From https://www.azul.com/eliminating-java-update-confusion/:

Oracle stated that their goal was to eliminate any functional differences between these two binaries. This will be complete with the release of JDK 11 in September. To achieve this goal, certain features (such as Flight Recorder and Mission Control) have been contributed to the OpenJDK project, whilst other features (like Java Web Start and JavaFX) are being removed from the Oracle JDK.

From http://www.oracle.com/technetwork/java/javase/eol-135779.htm...:

As announced in September 2017, with the OracleJDK and builds of Oracle OpenJDK being interchangeable for releases of Java SE 11 and later, [...]

fcurtson Do I need to pay for Java now?

The quote describes the situation before JDK 11. As of 11, performance of Oracle JDK and OpenJDK will be exactly the same. They'll sell support and certified LTS builds.

fcurtson Do I need to pay for Java now?

As of 11, a "closed source Oracle JVM" won't exist anymore, and there won't be any performance difference between Oracle JDK and OpenJDK.

fcurtson Do I need to pay for Java now?

I'd be surprised if the exact same codebase (as of Java 11) was 4x slower just because it's called OpenJDK.

fcurtson Do I need to pay for Java now?

You won't have to pay for using Oracle JDK 8u181 in prod, and you won't have to pay for using OpenJDK 11 and higher in prod.

You will have to pay for using Oracle JDK 11 and higher in prod, which is built from the exact same codebase as OpenJDK 11 but offers commercial support and long-term support (LTS) releases.

https://adoptopenjdk.net plans to ship LTS releases of OpenJDK for free.

fcurtson Do I need to pay for Java now?

That information is outdated. I'll say it once more: Oracle JDK 11 and OpenJDK 11 will be built from the exact same codebase.

https://mjg123.github.io/2018/05/25/Opening-Oracle-JDK-Featu...

fcurtson Do I need to pay for Java now?

> How is a license that puts no restrictions on usage "less business-friendly"?

Use of GPL licensed software is heavily restricted in many corporations, whether or not it has the Classpath Exception. Among other things, it's hard to verify/enforce that projects only link against files that have the Classpath Exception.

fcurtson Do I need to pay for Java now?

> TL;DR: Yes. You have to pay for Java now.

You have to pay for Oracle JDK but not for OpenJDK, which is just as good, except for a less business-friendly license, no commercial support, and no LTS builds straight from Oracle (instead you need to put your bets on adoptopenjdk.net).

> Quoting the article: «Oracle JDKs may only be used in production if you buy the commercial support.»

Yes, but the author forgets about this in the "What does the new release train mean to my company?" section and draws completely wrong conclusions.

> The alternative is using OpenJDK in production. We'll have to see how much the Oracle JDK and OpenJDK will diverge.

They will be identical. All of Oracle's JDK work, even LTS related work, will go into the OpenJDK repo.