https://discord.umbraco.com logo
Dislikes V17 Backoffice
# umbraco-chat
m
I personally do not like the new backoffice. It seems to be MUCH slower to use and a lot clunker then v13. I am forced to be on it right now as my projects have faced enough delays I don't want to deal with delaying again to either roll my own custom environemtn or figure out a more friendly CMS system. Once the project is launched and plan on sitting down with my developer and having a good long conversation on moving off of Umbraco due to what appera to be some core limitations as well as the IMO really bad v17 backoffice.
j
Hi, Jacob here from the CMS team. Curious to know if you have any insights you'd like to share? If you waved the proverbial magic wand, what would you change immediately?
p
@User May I suggest “Umbracians” instead of "guys"? We use gender inclusive language in this Discord 😀 ([read more](https://skrift.io/issues/happy-new-year-umbracians-let-s-start-our-conversations-friendly/))
m
Hi Jacob. The bigest issues I have, as it affects both me and my employees, is how slow the backoffice is to loads. It will randomly just decided it wants to take forever to load up. Sometimes it is just the main dashboard when first loging in, other times when we click on a document there will be a blank page for a few seconds before the properties load in. We can't resize images in the RTE. And a lot of small QoL things like if a required field is on a non-active tab but was not filled out when you try to publish you get an error message but nothing really indicating where the problem is unless that problem is on screen. The backoffice has come a long way since introduction but IMO you guys jumped the gun and took a way a very good, very established interface and replaced it with a subpar one that even now lacks some of the same functions and the old system. I know the speed issue has a discution thread and that the team is "always looking at ways to improve" and that image resize is waiting for the new RTE to ahve it as a core function in there product but the fact of the matter is that these core functions and issues exist in an LTE product designed to be used over v13 but does not live up to the abilites and stability of v13.
As for the core limitations two things are hitting me. The first I am not sure is actually a limitation but everytime I bring it up everyone basicly tells me that Umbraco is not sutied for it, the other once is doable in Umbraco but just not easily. 1. My site is a game information site so there are lots of pages for weapons, armor, items, food, quests, ect and they all interlink with eachother. A single game we are working on just over 50,000 pages for that information alone. This brings big concert for indexing and searching. While I have been told by some people who appear to work at Umbraco HQ that Umbraco can handel that many pages with no problems they do say that there might be indexing issues because of the large amount of data. Others have told me that Umbraco is not designed for these things. I really don't know, I do know I have ran into indexing problems back in v13 but those problems have been fixed in v17 bu we have not imported the information pages yet so time will tell on thise issue. 2: Dynamicly gereated and interactive pages seem to be out of scope for Umbraco. Some of the projects I want to build for my sites do not just present data but are user driven. For example, a crafting shoping list system. The user searches for the items they want to craft, the system looks a the recipies and then calculates the items needs and in what quantity. The user can then mark off what they collect as the collect them, and even save lists for feature usage. The only way I see of doing this is to create a doc type with no prooperties and wire all the code up in the back, but that means a single doc type for any and all interactive pages like this.
I had also forgoten to mention that the face password managers can cause problems with the backoffice "due to the shadow DOM" is really bad. Passwords managers are a great way to enhance account security on the net and to be like "yea, to make sure you dont have problems with umbraco uninstall your password manager"
j
Hi, thanks for getting back to me. Obviously, as you said, there are open issues for some of your concerns, others have been completed in recent versions. Not going to defend our decision to create a new backoffice, except than saying that a lot of people wanting to just 'have the old backoffice again' seem to forget all the quirks and pits making it unmaintainable essentially. We couldn't even move a single around without breaking an extension or package due to AngularJS's DOM-based scopes. The fact that we allowed to put in arbitrary JS and CSS all over the backoffice caused even more problems. Essentially, the old Backoffice had a reputation that any new even minor update would probably break things and you would have to wait at least until .1 or .2 of any new major version. Due to the tight coupling in all modules in the old Backoffice, and the fact we essentially couldn't lock away any services or restrict any UI in the Backoffice meant, that we could have never implemented anything like the new Extension API allowing developers to extend the UI in new, well-supported, and maintainable ways. All of that said, I haven't even mentioned the fact that AngularJS is EOL. We can't have our major product ship built on a framework that isn't supported anymore. I'm sure you can appreciate the fact that your auditors will not outright flag the CMS you intend to use.
The rich text editor issues, you mention, are tough ones. Yes, object resizing is not possible in Tiptap at the moment. Unfortunately, we cannot bring back official support for TinyMCE due to licensing. Tiny Inc. have an official package that adds support for TinyMCE, so you have the option to use it just like you did in version 13 and before. Just know that the MIT-version (version 6) is EOL (which is why we can't ship with it) and that you probably have to pay a license to upgrade to the newest version 8. Luckily, we will add more properties to edit in Tiptap, so that you can edit width and height on images. I believe that ships in version 17.4.
Lastly, password managers not being supported: I was actually under the impression that they worked now. Dashlane seemed to be the big culprit, but the latest reports I heard said that it now worked fine. Are you experiencing problems with password managers still?
Thanks for your words. I love hearing about the issues people experience with the product. Had a few more thoughts around what you said: Search & indexing - You are right, Examine will fail at some point given too much data. It's hard to tell when that happens. Ultimately, Examine is file based and so, limited by the filesystem. Did you know you can soon replace Examine? [Umbraco.Cms.Search](https://github.com/umbraco/Umbraco.Cms.Search) is currently in beta, but expect to ship soon including support for V17 and V18. It will allow you to plug in Elastic Search, which should be more than capable of carrying that amount of data. Dynamicly gereated and interactive pages - It seems to me like you are in a scenario, where you need to extend your application rather than force Umbraco CMS into a situation it's not suited for (square peg in round hole kind of scenario). You can add your own tables to the database, create repositories and controllers and handle the data like you would. You can do all of that whilst referencing Umbraco data. There are packages that add "custom data pickers" to Umbraco handling your data. [Contentment](https://marketplace.umbraco.com/package/umbraco.community.contentment) is one such example with its 'Data Picker'. a required field is on a non-active tab - Sounds like a [bug report](https://github.com/umbraco/Umbraco-CMS/issues) waiting to happen?
m
Search: I was not aware of the search replacement coming. That is good news. Dynamicly generated and interactive pages: I have not looked closly at these projects right now as they are probibily a year or two down the pipeline. Right now core data and information is the focus and these more specialty things are to follow that. The plan is to brin on an Umbraco pro on a contract to evaluate and propose solutions. If it can easly be done in Umbraco that would be great as there are features in the backoffice for page creation and mataince that make life easiser (picking links to other pages and them auto-updating if pages get moved, creating new pages without a system redeploy). I will admit the presumtion these projects are out of scope for umbraco is going of my limited knoledge of working with and inside umbraco. Required field: There was a bug report opened and was fixed in 16.3 but then broke later. Looking again they fixed it back in feb so I would have to push an update to prod to fix but prod is code frozen in favor of a rebuild we are in the middle of. The bug is at https://github.com/umbraco/Umbraco-CMS/issues/18096
I do understand why TinyMCE was removed and I had heard about the unmatainablity of the old backoffice. To me, and many others though I have seen a lot less talk about it since 17 relesed, the new backoffice was a big downgrade compaired to 13. Personaly when I was dealing with in 14-16 I understood these where STS and considered them more beta versions vs an acutal, production ready, release like 17. I don't know the structure of the teams at HQ and how issues are prioritized so I am not going to speculate on resorce allocation and priortization for getting the new backoffice feature equavalant to the old but the current issues my team and I are having are a big annoyanceI and have actually cost me money in downtime. The employee I have that had a PW manager installed on there system no longer needed it, it was a relic from an old contract they worked, and so have uninstalled it. Which looks to have resolved the issue of the the entire backoffice randomly refusing to load for him. I just started using my own password manager (proton) and so far have not noticed an issue but I don't often go into the backoffice as my resrouces are currently focused at an umbraco package I am building and then the rebuild of the site. So for now I have no data to point one way or the other. I do want to say that it was nice to see you not try to push for the subscription based plugin that Umbraco sells to manage outside data within the app (I think it is called like Data Builder or Data UI or something). I have actualy installed the TinyMCE plugin but with the performance issuse in the backoffice I am trying to keep to a minimum amount of plugins to reduce package calls and file downloads (which seems to be one of the leading theroies as to why the backend is slow).
j
> which seems to be one of the leading theroies as to why the backend is slow It is a very real thing, not just a theory. However, also hard in its nature to qualify since there are no absolute numbers saying "Ok, now it's good enough". Some people were talking about a 3 second time-to-input on any given document, but what is the configuration, how many blocks do you have? How many properties? Where is the site even hosted and what is your latency to that location? That's why we can only ever get as close as to a message like "We are continuously improving the speed of the Backoffice". Eventually, we will have to close off the issue, you mention, when we have fixed the things we are aware of (such as too many requests easily fixed by more bundling).
m
I do know how hard it can be to deal with things when the problem is "too slow", I spent 30 yars in IT, and the irony of me just saying it is slow is not lost on me. I wish I had the time to sit down and run timing test, screen caps, and more to provide more concreate data but the best I can say right now is that about half the time the page loads as expected and about half the time it feels like I am back on dial up and see litteral cunchs of the pages load, bigest impact are is when opening a content node, the main window will go blank and then after a noticable amount of time of blank ness the properties pop into existance oneish at a time. With the dynamic work a year or two out there is plenty of time for me to continue using Umbraco and watch it evolve. The only major close issue I see coming up right now is the search and that huge data import. Hopefuly it is not to much data at start and there is enough time to move to the new search function. And during all this time as I run into bugs I will continue opening reports and if I ever get more time on my hands to do deeper dive into the slowness issues I will be passing that on. There are many many aspects of Umbraco I like, I quite enjoyed v13, so I do want to help keep making it better where and when I can.
2 Views