#Alt-text

[jackgoodfellow]

Anyway, it turns out setting art goals according to what Julia from Drawfee can do is like setting swim goals according to what sea otters can do.

[animanightmate]

You can stop trying to find the best Kate Middleton joke, because it's here:

Source here.

[fayrobertsuk]

Occide, Regina

Made this for a friend who had someone give them the old "God makes no mistakes" line in response to them being trans.

(Image description in alt-text.)

[techniktagebuch]

12. Februar 2024

Wir werden also 5300 Bildbeschreibungen schreiben müssen

Bei der Vorbereitung der diesjährigen Techniktagebuch-als-Buch-Ausgabe zähle ich zum ersten Mal durch: Das Techniktagebuch enthält rund 5600 Bilder. Davon haben 300 eine Bildbeschreibung, die von Screenreadern vorgelesen werden kann. Bei 4700 steht im alt-Tag nur "image" und bei den dann noch fehlenden 600 vermutlich nicht mal das.

Ich habe so was Ähnliches erwartet, denn in den ersten sieben bis acht Jahren des Techniktagebuchs haben wir uns fast gar nicht um Bildbeschreibungstexte gekümmert. Schwache Entschuldigung: Der alte Tumblr-Editor machte einem das nicht so leicht, man musste die Bildbeschreibungen von Hand in der HTML-Ansicht des Beitrags einbauen. Wenn etwas so umständlich ist, suggeriert das auch, dass es eh nicht so wichtig sein kann. Aber eigentlich hätten wir es besser wissen müssen.

Ich weiß nicht, was bei den anderen Autor*innen des Techniktagebuchs die Gründe waren. Ich selbst bin vor allem deshalb nicht früher auf die Idee gekommen, weil ich erst seit wenigen Jahren selbst (früher bei Twitter, jetzt bei Mastodon) oft Bildbeschreibungen sehe. Dabei habe ich gemerkt, dass es erstens viele Menschen gibt, die sie sich wünschen und man sie zweitens auch als sehender Mensch gut brauchen kann, zum Beispiel wenn ich nicht verstehe, was auf einem Bild zu sehen ist. Oder wenn der Text auf einem Bild in der Ansicht auf dem Handy unlesbar klein ist (so wie in den Screenshots in diesem Beitrag hier). Oder wenn ich Hintergrundinformationen brauche, weil ich wegen meiner Prosopagnosie oder aus Ahnungslosigkeit abgebildete Personen nicht erkenne. Oder wenn ich mit dem Zug durch eine Funklochregion fahre und statt der Bilder nur die Bildbeschreibungen zu sehen bekomme.

Ich habe mit einem ungünstigen Verhältnis von Beschreibung zu Keine-Beschreibung gerechnet, aber dass dem Techniktagebuch so viele Beschreibungstexte fehlen, überrascht mich doch. Und es bedeutet, dass es sehr lange dauern wird, die Bildbeschreibungen nachträglich einzubauen. Da die Beiträge, die im alten Tumblr-Editor (vor 2023) geschrieben wurden, nur in diesem Editor bearbeitet werden können, heißt das, dass wir dazu jeden Beitrag öffnen, in die HTML-Ansicht umschalten und das alt-Tag mit der Bildbeschreibung an der richtigen Stelle einfügen müssen. Während man das Bild dann beschreibt, kann man es nicht sehen. Das ist im neuen Tumblr-Editor sogar genauso, denn da überlagert das Eingabefeld unpraktischerweise das Bild. Man braucht also in jedem Fall einen Second Screen, auf dem das Bild zu sehen ist. Ich mache das mit dem Handy.

Wegen dieser Umständlichkeit, und weil in letzter Zeit überall von großen Fortschritten in der Bildbeschreibung mit Hilfe von Bildanalyse und großen Sprachmodellen zu lesen ist, probiere ich heute mehrere Tools aus, mit denen das angeblich automatisch geht. Ich habe nicht aufwändig recherchiert, welche man da nimmt, es sind einfach die ersten paar Googletreffer zu dem Thema.

Ich teste mit dem als Letztes im Techniktagebuch verwendeten Bild aus diesem Beitrag von Oliver Laumann.

Oliver hat dazu die folgende Bildbeschreibung verfasst:

"Vierfarbkugelschreiber in Originalverpackung, Rückseite einer Postkarte mit dem Text 'Ich hab die Originalminen dringelassen, weil Original, aber hier sind neue. Grüße, Kathrin', kleiner Zip-Lock-Beutel mit vier verschiedenfarbigen Kugelschreiberminen".

Es ist sicher nicht ganz einfach, dieses Bild hilfreich und korrekt zu beschreiben, wenn man nicht Oliver Laumann ist, sondern eine Maschine. Aber die Schwierigkeit der Aufgabe scheint mir einigermaßen repräsentativ für die übrigen Bilder im Techniktagebuch.

Beim "Free AI Image Alt Text Generator" kann ich zwischen mehreren Sprachen auswählen:

Außerdem gibt es verschiedene Beschreibungsstile:

Ich wähle "Casual". Das Ergebnis ist wenig hilfreich:

Im Stil "Academic" wird es nicht besser:

Auch die anderen Beschreibungsstile führen nur zu minimalen Variationen von "Irgendwas mit Farbstiften, perfekt für kreative Projekte!"

Der "AI Alt Text Generator" bietet mir in der Testversion keine Sprachauswahl an und schlägt vor: "A package of colored pencils with a note on it." Immerhin ist die handschriftliche Notiz hier erkannt worden. Zum Verständnis des Beitrags trägt die Beschreibung aber nichts bei.

Der "Alt Text Generator" meint: "a box of pens and a package of paper".

Einige andere Tools verwenden offenbar intern die gleiche Technik und sagen ebenfalls "a box of pens and a package of paper".

Ich halte es für möglich, dass es kostenpflichtige Angebote gibt, die bessere Bildbeschreibungen erzeugen, finde aber im Rahmen meiner Suchgeduld nicht zu einer Möglichkeit, das gratis auszuprobieren. Und ich habe keine Lust, erst einen Testaccount anzulegen, nur um dann herauszufinden, dass die es auch nicht können.

Vielleicht sind Bildbeschreibungen nur eine Brückentechnologie, und bald werden Screenreader die Bilder selbst erkennen und beschreiben können. Hier denke ich mal einen Gedanken, denn ich eher selten denke, nämlich: Bildbeschreibung von einem Menschen, bevorzugt einem mit Wissen über das Bild, das nicht im Bild enthalten ist, wird immer besser bleiben als automatisch erzeugte Beschreibung. Gleichzeitig misstraue ich diesem Gedanken aus Prinzip, weil "Irgendeine Technik wird irgendwas niemals so gut können wie ein Mensch" eine Aussage ist, mit der man historisch oft unrecht hat. Zusammengefasst: Ich habe keine Ahnung. Aber diese Ahnungslosigkeit von 2024 ist hiermit festgehalten.

Jedenfalls werden wir 5600 Beschreibungen selbst ausdenken und einbauen müssen. Das klappt für die diesjährige Buchausgabe nicht mehr. Aber das Techniktagebuch ist ja ein Langzeitprojekt.

(Kathrin Passig)

[futurebird]

[blauespikmin]

A strange OMORI dream I had yesterday, visualized

ALT text under the cut

[Image Description -

The image has a vertical format: In upper half shows a starry cloudless sky. Central on the ground is Omori driving on a brown moped, emotionless. The ground he is driving on is a giant eye, like the eye of Something.

The black of the Something-ground develops into the lower half. It shows a large blood red body of water, with a smaller version of Omori on the side pushing a Sunny towards the lake. Sunny is just as emotionless and has his eyes closed. Both boys are proportionally more like the ingame sprites like this was a scene in the game. This scene is framed with thick black eyelashes and rather dark compared to the upper part.

End I.D.]

[unsergoldschatz]

Optimising image meta-data in Photoshop for improved Google search

Introduction

Content:

What are metadata?

Why is metadata important?

How to set metadata in Photoshop

Opening the File Info

Filling out the metadata

Which metadata are important for Google?

Description

Keywords

Copyright Notice

Author

Other relevant metadata

How Google indexes images

Additional tips for SEO optimization of images

Conclusion

FAQs

In the digital world, metadata plays a crucial role. They help search engines like Google understand the content and context of images. In this article, you will learn how to perfectly set metadata in Photoshop to be found SEO optimized by Google Images.

What are metadata?

Metadata is information that describes the content, origin, and other important aspects of a digital element. For images, metadata can contain information about the author, the camera used, the settings, and much more.

Why is metadata important?

Metadata is crucial for search engine optimization (SEO). They help search engines understand the content of an image and index it accordingly. Well-optimized metadata can help your images rank higher in search results.

How to set metadata in Photoshop

To set metadata in Photoshop, you first need to open the file info. To do this, go to "File" and select "File Info...". Here you can enter various types of metadata.

Opening the File Info

In Photoshop, you can access the file info by clicking on "File" in the top menu bar and then selecting "File Info...". A dialog window will open where you can enter the metadata.

Filling out the metadata

In the file info, you can enter various types of metadata, including description, keywords, copyright notice, and author. It is important that you fill out these fields as accurately and detailed as possible.

Which metadata are important for Google?

Google considers various types of metadata when it comes to indexing images. The most important are the description, the keywords, the copyright notice, and the author.

Description

The description should contain an accurate and detailed description of your image. It should contain relevant keywords that describe the content of the image and the context in which it is used.

Keywords

Keywords are individual words or phrases that describe your image. They should contain relevant keywords that users might enter in search engines to find an image like yours.

Copyright Notice

If you are the author of the image, enter your name or the name of your company here. This helps Google understand who created the image.

Author

Also enter your name or the name of your company here. This is another way for Google to identify the author of the image.

Other relevant metadata

There are also other types of metadata that you can enter in Photoshop, such as the camera settings that were used to create the image. These can be useful for certain types of search queries.

How Google indexes images

Google uses a variety of factors to determine how images are indexed and displayed in search results. These include the metadata, the context on the page where the image is displayed, and the alt text associated with the image.

Additional tips for SEO optimization of images

In addition to optimizing the metadata, there are other strategies you can use to optimize your images for search. These include using relevant and descriptive file names, optimizing the image size and quality, and using structured data.

Conclusion

Optimizing your images for search is an important step in increasing the visibility of your content in search engines. By carefully setting the metadata in Photoshop, you can ensure that your images are correctly indexed and displayed in search results by Google and other search engines.

FAQs

What are metadata? Metadata is information that describes the content, origin, and other important aspects of a digital element.

Why is metadataimportant? Metadata is crucial for search engine optimization (SEO). They help search engines understand the content of an image and index it accordingly.

How do I set metadata in Photoshop? To set metadata in Photoshop, you need to open the file info and fill out the appropriate fields.

Which metadata are important for Google? The most important metadata for Google are the description, the keywords, the copyright notice, and the author.

How does Google index images? Google uses a variety of factors to determine how images are indexed and displayed in search results. These include the metadata, the context on the page where the image is displayed, and the alt text associated with the image.

[rivertalesien]

Dear @staff,

There is one thing you could do that would make this site so much better: add an automatic alt-text function. If Facebook can do it, so can you. This site is virtually unusable for low-vision/blind folks. Accessibility brings more people to the yard, right? Bluesky is already way ahead of the curve on this. Don't be left behind.

Instead of losing yourselves in corporate quicksand, make the site usable. A little goodwill goes a long-ass way.

[graegrape]

i am once again reminded of the fact that toby fox is a genius at comedic pacing

[tophats-tea]

sliding scale of posts about Jesus being a carpenter who was nailed to a cross

[mangedog]

this is a way better model... you'll still get transphobic & intersexist drs of course but i prefer this to male / female or even having separate questions for gender & sex.

[we can't see the full form, but i'd suggest having a "something else" option and dominant hormone question too.]

[jackgoodfellow]

(Fujiko is dreaming of all the ways she is going to use her new boobs for evil. And she should! As a treat!)

[original comic by @silkentine]

[animanightmate]

I can’t believe I’m about to write this, and you may well have seen this story elsewhere but...

JK Rowling is currently fighting with the Mormons.

Okay, maybe not all of them, or even their official governing body, or however it works, and maybe fighting is a strong term, but...

Okay, let me start again.

As I understand the longer timeline, it goes:

JK Rowling’s Harry Potter books come out at the beginning of the century and some very hard-line religious types lose their shit because it’s a) popular, b) explicitly got witches in it, c) is for children. So they protest the books and their apparently subversive, Satanic messaging, ban the books, even burn the books. Other Christians deem the books a delightful metaphor, some citing the manner in which other Christian writers like C.S. Lewis used magical metaphor to champion the fight of good over evil. While the Church of Latter-Day Saints doesn’t proclaim an official stance on her work, they seem to be broadly supportive and enthusiastic.

Then, roughly 20 years later (blimey, that feels odd writing that down), the author starts becoming ever more heavily – and publicly – invested in a brand of feminism that deems trans women an inherent threat to cis women (especially lesbians), labels trans men a waste of perfectly good potential lesbians, and nonbinary people hilariously deluded (unless they’re AMAB, in which case they’re also categorised as being as dangerous as trans women). Intersex people are considered a statistical irrelevance.

I summarise very broadly here, of course.

She is generally careful in her language and in the explicit things she says in writing (and has, after all, historically donated to explicitly trans-inclusive women’s charities and LGBTQIA+ charities), but she repeatedly, and with increasing frequency, backs up those people and organisations who are much more direct in their condemnation of anything other than the cis condition (and likewise those who support a progression to wider inclusion of non-cis folk).

My understanding is that this is something some of a particular subset of more hard-line Christians could get behind (Rowling herself apparently defines as a Christian, which surprised me for some reason). They had her back on this particular stance, as it were. (I feel moved to say here that the vast majority of my actively Christian friends, including members of the clergy, are trans-inclusive, supportive allies, but maybe I’m experiencing a skewed sample...)

Then it gets even more interesting, because on Thursday Rowling tweets that people claiming she’s transphobic without supplying explicit enough evidence are akin to “...  when Joseph Smith found the golden plates and nobody else was allowed to look at them.” Direct quote there.

(image description in alt-text; please let me know if it’s missing! If you’re wondering why the text is red, it’s down to the Shinigami Eyes extension on my browser.)

And if you scroll a few responses down the thread, Mormons start saying “Now, steady on there, Jo, we’ve been big fans of yours for ages, but you’re dead wrong here, because people did see the plates, and this is frankly a false analogy. You clearly haven’t been reading up on your Mormon history, which is frankly disappointing. Why would you throw us under the bus like this?”

Which could be said to look a bit like

(alt-text should be in there - shout if not.)

The responses are a fascinating blend of vehement trans and trans-allied folk attempting to provide evidence, contemptuous anti-trans folk scoffing at such stances, aggrieved Mormons, amused/ bemused ex-Mormons, some rather aggressively atheist tweeters, and all blending in quite the mélange of sub-threads. They also include JKR expanding on her own post by saying that, having heard people in the replies telling her that the plates weren’t hidden, her research indicates that a few did, but, if it is a retraction, it isn’t exactly couched in the most fulsome terms, let’s say: “Eleven people claimed to have seen the plates, some of them related to Smith, but there's debate as to whether this was a metaphysical experience or they genuinely saw them. And one man was allowed to hold the box but not look inside it.”

I suspect that people will find themselves able to read a great deal of subtext in a number of directions into that message. I find myself wondering where that reading might lead them.

(I hadn’t even heard about these blessed plates until now, so, you know – it’s all definitely been educational!)

I’m honestly intrigued as to where this will go next. As I understand it, Rowling’s LDS fanbase is not small, and Mormons are starting to respond in articles and comments sections and dedicated boards in an interesting variety of ways.

I guess the only other thing I want to point out is that I can’t find the original redacted tweet – I’m assuming that either those more invested in anti-trans vehemence found the original (Twitter’s text search facility is pretty darned efficient, after all) and made things difficult for them, or the original poster was alerted to the fact that someone with nearly 14 million followers had raised awareness of it and pre-emptively removed it. The thing wasn’t exactly doing numbers, so it really makes you think about reach and exposure online, and how we can and need to take responsibility for that and our public wording.

[saszor]

Tutorial for drawing characters with cleft lip! Sorry that it's mostly unilateral-centric but it makes up the vast majority of resources and photos. Still tried to get tips for drawing bilateral clefts in though.

Keep in mind that this is an introductory drawing tutorial and has some generalizations in it, so not every “X is Z” statement will be true for Actual People 👍

If you draw any characters using this feel free to tag me!

[their-we-go]

and i'm going to.

[dasloddl]

ID in alt

we get it, it’s beans ON toast and ants would eat that, also we get it, ants aren’t that loyal to their queen, they kill her off if she’s out of line