August 9, 2013 at 18:40The phrase "reloaded on habrastorage" makes no sense
If before years the Habr site used third-party image display services, in connection with which the images of deceased services distorted the appearance and sometimes the meaning of old articles, now they’ve gotten hands and opportunities so that the site no longer depends on the arbitrariness of random hosting services - all images are automatically copied to habrastorage.org. This inconspicuous work on comments began earlier, in July, on the content of articles - recently, in early August (2013). You don’t even need to give examples - look at the HTML source code of any recent articles - in them you will no longer find domains with "left" pictures. Or better look at the code where the author reliably used third-party image hosting: habrahabr.ru/post/189474 , because most authors voluntarily use habrastorage.org.
Here is the case when the author preoccupied with writing pictures on habrastorage.org:
And here is the case when the author of the article used the usual third-party service (it doesn’t matter which is unknown):
The difference is that the robot did not add the autosaved image to the habrastorage.org domain, but to the habr.habrastorage.org subdomain. And that’s all that remains of the former connectivity and anarchy of the Internet.
Older articles (July and earlier) are not subject to change by the robot.
The result of the innovation for users has significantly more advantages than disadvantages. Now the pictures do not depend on the arbitrariness of third-party services, but the author, as before, when editing the article can change the picture to an updated one (the robot will immediately immediately write it to habrastorage.org cache again after editing).
There are other pluses. The minus is seen in two points: the author cannot manipulate the picture directly from his hosting and 2) - the author cannot set counters and publish articles with statistics on visits, as it was recently possible to do here: habrahabr.ru/post/186864 . This is almost the last article on pageview statistics (with graphs).
Such a minus described in a previous article is also not a minus and is easily solved by reference:
The use of external pictures made it possible to place at the end of the blog recording a button-picture combined with the counter of its clicks. Something like “Tweet! “150 people tweeted.” (Or with a meter of some other quantity. For example, “Donate money for the above project! - 320 050 rubles already collected.”) Now the picture will stop changing, so this technique is not possible at χαе.Because, firstly, clicking on a simple picture hadn’t done anything before, and the effect of the counter could be achieved only by a link that opens another picture-counter. But now - we are writing a link that opens a page on another site with a counter - we get the same counter.
Conclusion: you can now post pictures anywhere, if only they were read at least once. For example, on your Dropbox. After a few seconds of the demonstration, the robot itself will save the pictures from the article on the Habra website, and dropbox or Amazon hosting will be protected from the Habra effect and excessive spending of money on traffic for the author (the owner of the hosting with a fee for traffic). (The issue of quality, image shrinkage has not been investigated, but I will quote Boomburum's answer on this subject:
2) The permissible size of the uploaded image was increased from 800x1000 to 1920x1200 (or maybe more in height, I don’t remember) - this is enough to fill even the largest screenshot with 24 "monics. Someone will ask" what about retina "And it will be right, but we must understand that we are not doing photo hosting for full-size photos.
Pictures of higher resolution will now need to be shown only with links or, for example, links with pictures with previews. At the same time, the site with the help of the robot will itself make a medium-quality preview, up to about 1000 pixels in size. wide.)
An example of how to publish high-resolution images:
PS: links to third-party page resources are not yet replaced by saved copies.
