EEVblog Electronics Community Forum

My apologies for commenting to a thread which last saw activity almost exactly a year ago.

For information, Arduino has just migrated from SMF to Discourse and I don't think anybody is happy.

Part of the problem was managerial: apparently not even moderators or prolific and well-respected fixers-of-folks'-problems were given adequate opportunity to review the new platform. Another part is obviously "things have changed and I'm sufficiently busy that I resent having to put time into familiarising myself with the new stuff".

However there are also other problems: as mentioned in the post above pages aren't fully-populated and in fact scrolling upwards two or three messages generates network traffic as that portion of the page is reloaded.

Another underlying problem appears to be that- at least as used by Arduino- it uses some sort of "cloud-based" architecture. The result of that is that the server IP address changes regularly (i.e. every time you open a link in a new tab) which triggers protective software such as NoScript: if we're having to permit access to a large number of "cloudy" addresses I find myself what other services might be colocated on those same systems.

It is apparently argued that Discourse is open-source (as in fact is SMF), and the description on Wp reports that it uses PostgreSQL which is arguably more expandable than MySQL (and less beholden to Oracle et al.). But the impression I get is that the user-facing presentation and network problems are sufficiently serious to outweigh architectural advantages... unless Arduino were really up against some poorly-publicised SMF limitation.

MarkMLl

Haven't been there in a while, but noticed this on the topics I looked at the bottom -


 This topic will close 4 months after the last reply.

Was that there before?

If you need to save an endless page (perhaps for archival or just to read en bloc), I can recommend a Firefox extension called Singlefile:

https://addons.mozilla.org/en-GB/firefox/addon/single-file/

Unlike the normal page save, it puts everything in a single file (surprise) instead of a sub-folder with a file-per-resource. You can annotate, save just a selection, etc.  I use it all the time to clip to a notes database in preference to using the built-in clipper.

Relevant to Discourse, it captures the entire endlessly-scrolled page, kind of like the 'print' option here. Unfortunately, it doesn't capture image pastes, presumably because the image is not actually on the page but only shown if you click it.

Edit: The sample I took when trying this out is attached, in case anyone wants to see what's possible with a single click. Completely unedited or otherwise messed with.

In the time since this thread was opened I've encountered several Discourse based forums, some I tried to like because I liked the content but ultimately I just hate it and abandoned all of them. It's just so... bad, there is so much wrong with the experience that it's hard to even put it into words. The fact that some people actually like Discourse suggests to me that some peoples minds are fundamentally wired differently than mine. At this point I've just decided that I refuse to use Discourse, I will not join any that are on it and will abandon any that switch to it, it just isn't worth the aggravation.

Quote from: james_s on April 25, 2021, 05:45:43 pm

In the time since this thread was opened I've encountered several Discourse based forums, some I tried to like because I liked the content but ultimately I just hate it and abandoned all of them. It's just so... bad, there is so much wrong with the experience that it's hard to even put it into words. The fact that some people actually like Discourse suggests to me that some peoples minds are fundamentally wired differently than mine. At this point I've just decided that I refuse to use Discourse, I will not join any that are on it and will abandon any that switch to it, it just isn't worth the aggravation.

No matter how truly terrible, God-awfully bad something is, there will always be a few people that like it!

The developers mothers don't count.

They must be doing *something* right based on the way it's spreading everywhere, but damned if I know what. Maybe it's just slick marketing, or the social media junkies that I can't begin to relate to.

More whitespace and mobile "fashionality" ??

P.S.  Where am I, and why am I in this handbasket??!

I hate mailing lists, I already get enough spam as it is without getting dozens of emails from some mailing list. IRC was cool back in the 90s when I was a teenager but I'm not really a fan of online chatting in general. Forums like this work really well for me, although it would be convenient if there was some sort of archive or a way to export threads of interest. Bottom line there I guess is nothing lasts forever, especially nothing on the internet, somebody has to maintain it or it eventually returns to that from which it came. 

...no way to archive the site as it is so it could be searched on something like the internet archive.

wget?

Hopefully, if Dave ever gets tired of running the EEVBlog, it will be handed on as a going concern to "someone or something"...   

Hmm, so me up all last night projectile hurling into the bowl may not have been a myserty virus... I suspect I'd better watch my back, something is afoot.... 

Quote from: aandrew on April 25, 2021, 11:21:22 pm

...no way to archive the site as it is so it could be searched on something like the internet archive.

wget?

wget works well for static sites where every document has a unique URL, and there's a "root" document with links in a tree structure. It's much less use for "modern" websites where there is no explicit tree, and locations are either identified like

https://www.eevblog.com/forum/index.php?action=unreadreplies;start=0

or like

https://www.eevblog.com/forum/chat/migrating-the-forum-to-discourse/msg3557316/#msg3557316

It's no use at all where a page is identified solely by a site name, and all content is generated by inline scripting

https://www.ebay.co.uk/myb/PurchaseHistory

In principle, it's possible to scrape content by generating every possible message number and sending a wget request for that specific page. Both SMF and Discourse appear to be amenable to that to some extent, and I've done it with Hackaday in the past to get a complete content list (I was trying to organise my reading, it's a useful site but I'm 18 months behind on it).

@aandrew mentioned the possibility of converting it to a mailing list, or IRC. I'd like to comment on that, and think it might be worth doing so in this message since anybody who isn't a comms nerd has already given up on it :-)

Mailing lists and Usenet-style discussion groups have the useful property that every message has a unique ID that in principle is generated when the message is first created and never changed. That's why clever software can thread messages and subtopics intelligently, which makes using a large and busy discussion group much easier.

Many email clients store messages in a flat file, which makes sorting and manipulation (i.e. "retrieve the message with this ID" difficult).

Some email clients and most discussion group clients and servers store messages numbered sequentially. Looking "under the hood", client software can rapidly locate a message using that sequential number but not by the original message ID. A major problem here is that a user can post a link to a message stored on his client system or local server, but that link usually contains the sequential number rather than the ID so is useless to anybody else.

In the UK there is a conferencing system called CIX, which at one time was heavily used by tech journalists and the likes. This had limited direct compatibility with mail/Usenet, BUT had the very important property that the unique message ID was guaranteed and any user could generate a message link containing it which was immediately useful to all other users and to the handful of offline readers (notably AMEOL, "A Most Excellent OffLine reader") which various people wrote to support it.

I'd add here that MediaWiki, like SMF, stores pages in a MySQL database and that to some extent that can be backed up and manipulated. Discourse's choice is apparently PostgreSQL, which is probably superior in its scalability etc. and in any event isn't beholden to Oracle.

Irrespective of backend, IRC messages carry no metadata and can't be meaningfully threaded. Good fun perhaps and OK for rapid-fire questions where a limited number of people are trying to talk at the same time, but much less generally useful.

The bottom line is that for archival and forking purposes, there's no alternative to getting hold of the raw messagebase and then seeing what server software can be used to present it. I've known various people who've attempted to generalise this, and have done a limited amount myself.

I'd like to emphasise that I am in no way suggesting a "palace coup" against Dave. SMF's presentation and Dave's site management are entirely adequate for what's going on here, although as a former CIX/AMEOL user I'd say that in principle the message threading could be improved. However if he did have any comments about scaleability etc. I think we'd all be very interested to hear them.

MarkMLl

Well the only two sites I frequent that are based in Discourse are the Rocky Linux Foundation one - the spin after the CentOS had changed to CentOS Steam - https://arstechnica.com/gadgets/2020/12/centos-shifts-from-red-hat-unbranded-to-red-hat-beta/ - that is a very small forum, and the Level1techs that looks like it is Discourse but can be something similar that is huge and with lots of technical info and activity.

My opinions are meh, I don't mind. It is true that "dinosaurs" like SMF, phpBB and vBulletin simply work and have a lot of development years plus a wide implementation on the WWW. But is also true that their layout is dated compared with the new trends, specially for mobile.

Discourse and other similar solutions - one proprietary that comes to my mind is reddit - are definitely more modern, nicely looking and support mobile and tablets natively or via an APP. Although the endless page load is very taxing on the memory of a PC. With 1 tab open on Firefox and 30 subreddits subscribed, some low traffic all day and others that have new messages each 10 sec in high traffic time, and uBlock extension working to not show the ADS, I can scroll until 1 day before (24 hours) and I start to note a slowdown on Firefox and 8GB of RAM consumed just on that tab.

Plus the internet points that most of this kind of solutions, a la Youtube, Facebook and Twitter Likes/Dislikes (that I despise, specially Facebook and Twitter that are mainly a Cesspool) that makes what is popular on top, most of the times as the laziest attempt possible, while good technical/informative posts are relegated to the bottom of the barrel because "they are boring" (the attention span of the new generation in full effect). For that reason I have reddit with the New option activated, that means that I see the posts chronologically by publish time and not by votes, like a normal forum is.

The reality is that both solutions fit their role, both have limitations. I think that it should be the kind of "don't change the winning team". Changing just for the fact of being "modern", "in", "trendy", etc means jack shit and will alienate a lot of current users, some who prefer the simplicity of how things are and others that simply don't want to learn a new way because how currently is works and don't need to be changed - and the human being is very reticent to change their way of doing things just for the sake of changing when both get the same end result.

Could look better, yes it could. But will it really matter when in reality we are here for the info and the exchange of knowledge, not for collecting badges or making a competition of who is the most popular. To me even the post count should be something that should be hidden or only count in certain technical subsections of the forum. I frequented back in the day a forum - the XtremeSystems Forums - that the counter was disable for threads created on the off-topic subsection that was the one with most traffic per day. Not only that but also because the Buy/Sell/Trade Section of the forum was blocked to users with less than 50 posts, so that way they kinda prevented the type of user who registers just to sell stuff.

But as also mentioned, all the info of the forum should be mirrored (and I'm sure that Dave done it already, mind it, this is not a critique to anyone, on the opposite) in case something very bad happens and all this info is lost. We already lost a lot of old discussions and tons of valuable info when Yahoo Groups closed and warn that all the info in the 18 years since they started was going to be purged, having the users to run against the time to archive years of valuable info.

I feel fine here, even if the "house" is not as modern as the other ones in the end of the road. The moderation and administration are spot on, not too much but also not too lenient. Even if sometimes the users are a little hard to talk with.  Most of the times I'm afraid of commenting because there is people here with heck more knowledge about something than me but with bad temper, but as an IT guy is the same, no time for bullshit and stupid questions so I understand perfectly well. That makes me to really select what I can comment and will not get eaten alive by any stupid thing I must say. But enough, that is my 2 cents.

Quote from: Black Phoenix on April 26, 2021, 03:48:28 pm

[...]  We already lost a lot of old discussions and tons of valuable info when Yahoo Groups closed and warn that all the info in the 18 years since they started was going to be purged, having the users to run against the time to archive years of valuable info. [...]

Perfect example of why you shouldn't trust cloud services to be long term stable...

Some clever peeps have run their feet off feeding the archive sites with this content. The sword of Damocles hangs over every platform. It's a shame that the archive sites can't be more interactive rather than static and frozen in time.

I snipped most of your message but I need to ask: did you write it on a smartphone, or did you use a computer?

Because I think that the vast majority of this forum's users access it using a standard web browser on a standard computer, not through an app or mobile browser on a smartphone.

You simply can't write a detailed response with, perhaps, graphics and images on a smartphone.

Anyway, one wonders whether the Discourse developers bothered to submit their product to actual usability testing. Actually, I don't wonder. Because if they did, their software wouldn't suck.

I use both the Smartphone, the Laptop and a Tablet. It depends of the situation, were I am and what is close by. Regarding writing an message with image links (IMG) and text on a Smartphone, is possible but time consuming, but I've done sometimes. But yes, most of the time I access this forum via PC, so it doesn't bother me that much.