Posts by Charlie Clark

I'm not sure that the syntax change will stop people writing code to avoid nested if statements.

DSLs are probably the way to go for the small group of people (relatively) who need this.

Python has a history of introducing syntactic changes only to reverse them later (backticks, print >>, map, reduce, etc.). Some stuff stays because it is genuinely useful but this can take years of refinement before it becomes standard. My initial reaction to this change is that it is, unfortunately, yet more special use syntax forced on the rest of us. It certainly won't aid readability and the improvements are marginal. But let's see what the take up is in the next few years.

Re: Yes, and...

The sort algorithm, which has since become a reference model, would suggest you don't know what you're talking about.

In previous discussions that were similar Guido tended to follow the argument that to do nothing would cause the least problem. This is exactly the kind of the syntax that, because it's occasionally useful, it gets used all over the place where not only it isn't useful but in fact downright confusing, aka dangerous.

Seeing as it's backwards incompatible

Re: Can you configure a database as case insensitive now???

Searching and sorting are not the same so I'm not quite sure what the issue. In general, for columns you should normalise your data for whatever processing you're going to do. Case-sensitive sorts are almost always going to be fastest and you can define pro-table or pro-column collations and you can even do this pro query. See the docs for more information. The docs aren't perfect but that was the first page that popped up when I searched for collation, though as long as you don't use the LOCALE_C you shouldn't have too much trouble.

For searching, there are worlds of difference between using the builtin LIKE or REGEX and the text search extensions. If you do have data that is case sensitive but want to do case insensitive search (you have "JOBS", "jobs" and "job" in your data), then you really should normalise your data and/or query, or at least use a relevant index.

For full-text searches case sensitivity is the least of your worries!

If I see something like query speed compared then I generally think, someone hasn't thought enough about their indices, or tried to use the analyzer. Now, while there is definitely scope for the Postgres analyzer tools to improve so that you know what to look for, that doesn't mean the database is slow. If something is "mission critical", be prepared to spend some money on external support to get the design right.

What do machine learning and "AI" have to do with a relational database? Please take your buzzwords with you and close the door on the way out.

It sounds to me like you're describing MySQL. Postgres has always had excellent transactional and parallel features.

You probably don't mean to say it, but your post implies that writing extensions for Postgres isn't easy. This is, of course, far from the truth.

I'm struggling to think of features that SQL Server has that Postgres doesn't come with or have extensions for. Maybe management functions? In other areas Postgres seems to be leading the pack.

For really high-end deployments, EnterpriseDB's version comes with support for Oracle's backend which makes migration of some large installs feasible.

The text search engine is awesome.

The whole Debian on the Gemini was a huge distraction for Planet for a few vocal users. Because the device uses a Mediatek SoC getting proper support on Linux was always going to be difficult. For those who want to go that route, Sailfish is the better option. But personally, I think Android with a good SSH client is sufficient for most sys admin work that I'm likely to want to do on the move.

Re: Software

I might indeed switch to Sailfish at some point, but it's still not a ringing endorsement of Planet's approach. They could have gone with AOSP from the start, except that this doesn't get them the Gapps that many people consider essential.

Re: Software

Unfortunately, the software is the Achilles heel of Planet's offerings. The hardware Gremlins have been resolved but firmware updates are few and far between and long-standing bugs just haven't been fixed. The device itself is immensely practical. You really need a table to use it properly but it does then let you do a lot of stuff: I was able to demonstrate PyDroid on it recently, albeit without using the native HDMI out, which for some reason refused to work. USB-C on one side is charging only, on the other output only. Sound is tinny but bearable. I only got a replacement Germay keyboard after Andrew Orlowski intervened. This is all okay for a device that you can always takes with you and works.

But the software side is poor:

• Android 8.1.0 from Decemeber 2018
• if WLAN is dropped, it will not reconnect but drain the battery in a couple of hours
• keyboard will randomly go into CAPS
• quite a few apps (IPlayer Radio) just crash
• the custom version of K9 is quite frankly awful

Basically, looks like Gemini users have been frozen out so that the Cosmo can happen. I knew what I was getting into when I signed up, but still disappointing to see the after-sales service fail so badly.

Re: Well that ruling has a timespan of about 30 days in the UK

There won't be a confidence vote until Corbyn agrees that someone else should be caretaker Prime Minister. His minders won't let him pull another December stunt.

Trying to repeal the Benn Act on any grounds would almost certainly be quashed by an injunction. The Supreme Court has asserted the primacy of Parliament and it's difficult to see even the High Court trying go against that judgement.

Bojo could resign but the fixed term act doesn't mean this will lead to a new election. Cf. resignations of Cameron and May. May eroded an already slim majority through bloody-minded attrition but Bojo just threw it away in a fit of entitled pique.

PMQs this afternoon with a decidedly C-team: Raab and Abbot

Re: Well that ruling has a timespan of about 30 days in the UK

Parliament also has the nuclear button of revoking Article 50.

Re: Well that ruling has a timespan of about 30 days in the UK

Who hasn't done what?

I think the government might be planning to break and subsequently challenge the law in which case, given the Supreme Court's recent judgement, it's likely to find itself subject to an injunction and likely to be held in contempt. And then there are still plans to prorogue for a Queen's speech…

The B-team has already demonstrated it doesn't have the best legal team and Bojo's declaiming that he won't comply won't help him in court either.

Pity the fools who have put their trust in these idiots.

Re: That was nice

Well, we now have a ruling for the lower courts to follow. A couple of summary judgements should be sufficient to get most people to start to comply

Re: Well that ruling has a timespan of about 30 days in the UK

As things stand Parliament has passed a law that the UK will not leave on October 31st without an agreement including the continued observation of GDPR.

Re: That was nice

Parameters are particularly susceptible to interception which is why this kind of session handling has been frowned upon for years.

Re: That was nice

How's this supposed to work with a stateless protocol like http? The cookie is required to map browser requests to the server session.

Re: That was nice

The ruling is quite clear: consent must be explicitly given so such messages are invalid. This shouldn't really surprise any once as it has consistently been the position of the courts.

However, GDPR is currently being revised and it wouldn't surprise me if some cookies become legitimate without consent as long as there is a list of them with their purpose and lifespan. It's not as if this hard to do, though I've seen multiple, ahem, "developers" fail to provide a correct list and explantion of a website's cookies. Note, that failure to do this correctly could go beyond GDPR and enter the realms of fraud.

The notice that somehow the BBC News homepage is better when you're logged in is, indeed, pretty Orwellian. The Charter almost explicitly mandates against this because the desire is to provide a universal service.

Re: Why is it even an issue on SE?

I am a moderator but I don't have time for it, which tells us all we need to know about crowd-sourced moderation. My time is devoted to answering questions on my software.

I like SE. I have found lots of usual information on it and I hope I've helped others and I think that the software gets a lot of things right that other forums have failed and the credit system isn't completely flawed.

But that doesn't mean that there isn't a lot of crap there.

Re: Is this just an English thing ?

If you think our social constructions (girls wear pink, boys wear blue, for example) are biological

I don't, which is why I didn't say it.: those are roles. FWIW pink used to be considered the male colour.

As someone else has pointed out, it can be helpful to use "sex" for biology and "gender" for grammar, but I did point out that the two should not to be considered the same, though they are generally correlated for animals in many languages.

But I don't agree that gender can be self-assigned and I'm not at all a fan of the attempt to normalise sex-change medicine. I won't necessarily go out of my way to point this out.

But by and large, I find these CoCs are little more than a conscience-salving by an IT industry that has allowed itself to be cowed by some well-organised lobbying and American tort law. And it's always easier to sign up to a CoC than actually do anything significantly different.

Re: Is this just an English thing ?

They to refer to unspecified individuals has been used for hundreds of years. Gender is a fact of biology that will survive any linguistic fashions. It's not the same as sexual orientation and distinct from any grammatical representation. The rules for any particular language are arrived at by consensus of the speakers, which means that changes by diktat are rarely successful for anything other than spelling, because the speakers have an innate sense of what is right or wrong for the language.

Re: Email?

In general, you're absolutely right: state-issued identities are really useful in many situations as they largely solve the problem of trust.

But, I think the problem is that some of the implementations are leaky, because they are supposed to be used for things other than simply signing documents electronically. Unfortunately, I think we have to assume that if there is opportunity to collect metadata (party A signs an agreement with party B) then the opportunity will be used and the data at some point will be exploited.

In the UK, of course, this would never be open to government abuse as the service would be awarded to the biggest donor cheapest contractor most reliable contractor (oxymoron when it comes to IT outsourcing).

Re: Email?

S/MIME is not considered to be that good, which is why in many goverment sanctioned situations, it's not acceptable.

PGP is a better approach but suffers from the problem of all decentralised systems: who do you trust? Though, it's not as if CAs haven't been compromised in the past.

Re: Email?

In this case I would expect examination of the headers to be important than a signature. Most UK companies should be using TLS in which provenance can be established.

The wording of the e-mail makes fairly unequivocal reference to agreement and does not look like an invitation to treat or to solicit an enhanced offer.

Re: "while working on Rupert Murdoch's organ."

He was sacked for working at The Times when he could no longer cover up the shit he was making up. Of course, it's exactly the same kind of shit that The Telegraph laps up, which is why, even as Prime Minister, who might be considered to have more important things to do, he continues to be paid to write opinion pieces.