Hello. I’m a relatively new discourse user. Here’s a thread I created earlier about the issue I was having:
I don’t know if this is a bug or not. The feature I’m asking for is keeping quote levels accurate when quoting partial text from someone. For example,
using this: A History of Pizza | History Today :
Pizza is the world’s favourite fast food. We eat it everywhere – at home, in restaurants, on street corners. Some three billion pizzas are sold each year in the United States alone, an average of 46 slices per person. But the story of how the humble pizza came to enjoy such global dominance reveals much about the history of migration, economics and technological change.
In general. I guess so, yeah. Here’s a typical scenario I’ve run into which I’m trying to address:
Pizza is the world’s favourite fast food. We eat it everywhere – at home, in restaurants, on street corners. Some three billion pizzas are sold each year in the United States alone, an average of 46 slices per person. But the story of how the humble pizza came to enjoy such global dominance reveals much about the history of migration, economics and technological change.
You just want to address/disuss/point out the part about three billion pizzas so you do this:
So what’s you’re asking is that the quote function should include the context, by “extending” the beginning of the selection (without actually copying the previous text) recursively until it reaches the beginning of context.
I.e. a selected text can be inside <b>, itself inside a <a>, itself inside a quote, itself inside a parent quote, and so on.
The issue you describe is how the quote works overall: if I select only part of your A History of Pizza | History Today link, it will quote the text but omit the link (and the quote).
It seems a good idea to include this kind of context, but I feel like this would be complicated to do
I would dislike such action. What I select is what I want to quote, and software shound never change that. It is always bad idea when a software, aka. a dev, thinks it knows better than a user and starts fix things. For such works we have Microsoft and Windows
This is a matter of what should be preferable between the two because we can’t have the best of both worlds. Either we quote the context as well, which might be valuable, or we omit the context, which might be at a loss.
I’d say that most of the time I quote, it would have been better with the whole context of the selection.
Pizza is the world’s favourite fast food. We eat it everywhere – at home, in restaurants, on street corners. Some three billion pizzas are sold each year in the United States alone, an average of 46 slices per person. But the story of how the humble pizza came to enjoy such global dominance reveals much about the history of migration, economics and technological change.
Now let’s say you or someone else wants to quote the thing I shared for a discussion. If you quote directly from me and quote just part of it, like I mentioned, something like the billion pizzas part it would come out like this when pressing quote:
If someone isn’t paying attention when using discourse it shows me as the one saying that. Thats technically a misquote. I don’t know the technical feasibility of this but it should quote it with a > to show that I am not the one saying it.
The problem with the current functionality is it leads to misquotes. I’ve multiple times had people post a quote on my forum that says I wrote text, but the text is actually from some book I quoted. If the quote level was correct, then people would see the attribution wasn’t quoted rather than thinking I wrote the text.
I think it should be safe for non-technical, non-precise users to select any text, click quote, and get a quote that isn’t clearly wrong.
Email lists handle this better. Observe me selecting quote level 3 text and then hitting reply in Mac Mail:
Expecting users to always quote the attribution in addition to the quote doesn’t work well. Users may not know to do that. They may not want to clutter up their post with extra text. They may be only quoting the middle of the quote, so to get it right they’d have to manually edit out text in between the attribution and the part they’re quoting and type an ellipsis. They may be quoting multiple sections of the quote to reply to separately but they wouldn’t want to include attribution multiple times in their post.
I do think there are also issues when you e.g. quote a phrase that was bolded and get a non-bold quote because you didn’t quote the start or end of the bolded text. But this tends to be much more visually obvious to posters in the preview than a missing quote level inside a blockquote, and the harm of errors here is smaller than with incorrect quotation attributions.
I think that however you handle this stuff in general, quote levels and attributions in particular should never be wrong when users don’t manually screw them up.
I selected a bold word but that is not what I got.
Why? Because under the hood the original looks like before **Bold sentence with words.** after but that is not displayed to the user doing the quoting. They saw rich text, selected a bold word in rich text, and expected it to come out bold. They didn’t see themselves as selecting a section of markdown text that didn’t include asterisks.
You call this the software not changing the selection and just being literal. But these results only make sense in terms of rich text. In terms of the underlying markdown, the software changed it on me. My selection didn’t include the closing **, but the forum software added that in.
If you want people to quote from raw markdown, display it to them and let them select it directly to quote, I guess. Having them select rich text but get quotes (partly) based on markdown is confusing and leads to results users don’t intend.
They’re the same issue. Are you saying you think the “with” should be bolded in my first example to you? That’s the same thing as including the > from the underlying markdown that applied to the selected text (in terms of the displayed rich text) but was not included in the selection (in terms of underlying raw markdown).
And when a user highlights some text, hits the “quote” button, and then writes a reply under it, that is not user error. That is not an issue of people can’t quote right. They used the provided user interface in a perfectly reasonable way. But it sometimes results in incorrect quotations.
First one is question of users who don’t choose at all those lines they should to get quote they would like to quote. Same thing than left half of a post pasted and still want to get unpasted too.
Second one is a glitch where quote doesn’t markdown, if it is the only part quoted. Text itself is quoted, though.
The user selects text, displayed in rich text, where he can clearly see it’s part of a block quote (same as seeing it’s bold). Then he quotes it and the markdown (> in this case, rather than **) is omitted because he didn’t select the place where, in the raw markdown, that > is. It’s quoted as plain text, not quoted text, even though the thing the user selected looked liked quoted text to him. So then the user falsely presents Dickens’ text as if Elliot wrote it. That’s the problem this thread is about.
The correct quote that I think Discourse should do is like this with the increased quote level, like just like how email clients worked decades ago:
I don’t think that including the context implies that the software guesses what amount of text should be quoted.
The text is still what we selected, no less no more. Adding context doesn’t mean it adds text. It adds formatting, and only formatting.
