I agree that they accumulate, but not necessarily that they don't break. I've recently recycled a 600e (working), a T30 (working), and a T61 (motherboard dead). There was a massive build quality difference between the (brilliantly solid) 600e and the T61. I'm hanging on to an x61s for the time being, since it still works and it's quite small.
I still love the design ascetic of these laptops and I wouldn't be surprised to find myself buying one in the future.
The T61 was the first laptop I bought and it completely put me off of Lenovo. Complete piece of shit -- broke multiple times even after I had them repair it and I even waited like 4 months for my order to arrive because they had some sort of manufacturing problems.
I must say, having switched from Google OpenID 2.0 to Google OAuth2.0 for authentication on our site I raised an eyebrow when I belatedly read on Homakov's 2013 blog[1] that OAuth2.0 was never intended for authentication, only authorization for certain discrete resources. At least with Stack Exchange doing it I'm in good company, but - if Egor is correct - it raises the question what's really a good choice for authentication nowadays.
Mozilla Persona is the right approach to authentication. Unfortunately, Mozilla sucks more at marketing than Microsoft did in the late 90s and Persona got near null interest from developers that weren't already concerned by the very real dangers of OAuth and the uprise of Facebook/Twitter authentication.
Persona is still there. It's still maintained, despite getting no funding, but that doesn't mean developers can't pick it up. If you are interested in authentication, learn about it. Implement it. Support it. Convert people.
It's true, OAuth 2.0 alone isn't intended to be used for authentication. Think of it this way: An OAuth access token is often described as being analogous to a valet key — anybody bearing the valet key can unlock whatever it is that the valet key unlocks, but this says nothing about the bearer of the valet key.
Google now uses an OAuth 2.0 extension called OpenID Connect. This introduces an entity that's analogous to a referral letter [1], the ID token. It's basically a little string of encoded (possibly encrypted, possibly signed) JSON containing 'claims' about the authentication state of the end user. The client application can then validate that token to confirm to its satisfaction that the authentication happened for some particular user — and that the ID token was created for it and not some other application.
You get all the benefits of a mature and very full-featured RMDBS.
Neo4j is a great database, and I use it myself for a project that involves lots of deep hierarchies. But it has a fairly sparse feature-set where schema enforcement and data integrity checking are concerned; you basically have to add all that stuff yourself at the application level which can amount to a lot of work.
To second this, if you're using multiple languages to write to your database, you almost certainly want that schema enforcement at the DB level. We decided not to go with Neo as our canonical database for exactly that reason. Its main strength is probably as a follower or slave to a RDBMS or log file, which is how it seems to be used in enterprise - when you want to do graph-based analytics, you bulk-load a snapshot of your non-graph-stored DB into Neo, then run read-only workloads.
I think it's fair to hope for a more customizable Chromium, since customizability is one of the things Opera was known for. For me personally, things like a bookmarks sidebar and more flexible tab positioning would help tempt me to switch back to Opera from Firefox.
> it's not possible to unambiguously segment the text into distinct letters (which is a necessary first step in any OCR engine that I'm aware of)
In my experience, the ability to handle overlapping letters (which is very common on type-written text and professionally typeset material) is one of the key things that separate the relatively lightweight OCRs (like Ocrad and GOCR) from the big complicated ones (Tesseract, Cuneiform, Abbyy etc). Whitespace character segmentation cannot be taken for granted if you want to do any useful OCR of "historical" material.
I might actually do that. I see that people are persistent in believing it's hard, maybe after making these claims I have an obligation to back them up.
That's not the case in my experience. Anecdotal, but I worked at a company where the brunt of the downturn layoffs fell disproportionately on people who needed work Visas, which have become much more difficult to justify in the current climate. AFAIK nothing's changed in that respect.
