Every bullet point says "Missing...", but I feel like the main issue about Jira is that it has waaay to many features, and that it's super slow!
The only bullet point I do agree with, is that Jira is missing an easy (or IMHO a minimalist / simple) interface. The tool is a mess, and the people using it are probably also using it in the wrong way, making the whole situation way worse.
Don't get me wrong, I like Jira to some extent, but its performance and messy UI are really making my life harder rather than simpler.
A former PM prioritized issues in the backlog by making a ticket and giving it a name like "Groom these next" or "Prioritized for January release", placing it in the backlog, then placing the tickets to be prioritized above it
They also hid bugs from being counted in reports by marking them as stories but that's another conversation
The interface used to be simpler and less in the way of getting things done. Then they added personas, UX and tried to "help" the user. You now need more clicks to accomplish the same stuff while information is more hidden than before also.
There's no way to cater every possible workflow without feature bloat. However, the space is ripe for simpler, more accessible solutions that supports work across multiple projects.
It's not for engineers. Or anyone in particular. They're building to check off boxes in order to sell top down to C-level execs that never touch the stuff.
Atlassian products are all like this, and each one is a horrible trash fire.
Confluence is the worst wiki product I've ever used. Gross syntax, slow, clunky plugin system, awkward permissions. Install MediaWiki and be done with it.
Bitbucket is slow and feels like it's from the 90s. It doesn't seem to do multithreading well - sometimes force pushed branch rebases pick up a new, unrelated author. A bug that hasn't been fixed in years, and only one of many. How many times will merging not actually work?
Jira is a special kind of hell. It's like every team needs their own embedded Atlassian solutions engineer to get it into a usable state. Not an EM or PM, but a specially dedicated role just to deal with Jira nonsense. Why even bother?
Atlassian is like Salesforce for business process, but if it were written by all-remote temp contractors churning through Jira tickets themselves.
Agree with this. My hypothesis is that they’re just ridiculously slow at adopting newer UX principles. The product was successful initially (good for them) and it was just never made much better.
The horrible response times and user hostile UI is shocking, especially if you look at their competitors. It gets in your way all the fucking time. But because they’re embedded in enterprises they don’t need to change much. It’s a big mess.
> Confluence is the worst wiki product I've ever used.
BUT... the direct integration with the ticket system - highlight a list, and make linked tickets (which keep the status info 'live' in the confluence doc) is pretty powerful, and allows a lot of people to keep their business docs, and get status updates without ever having to learn other parts of the system.
I was feeling the same way. Maybe we can get a Steve Jobs of agile development and create the "Iphone of Software Development Management Software". What I (drunkenly) mean is that the customer might not always know best when it comes to UX, especially when your early adopters/feedback providers are human management nerds.
If you want something that is simple use kanban-board like trello, github, an excel spreadsheet or whatever.
The complexity comes from the demand of Jira's paying professional customers.
Eventually Jira will grow too complex and bane the way for simpler tools probably of a different paradigm. Much like how Jira replaced things like HP Quality Center or the tools of Rational Unified Process.
I think you are right: the SLOWNESS is a killer. In fact, I think it is THE show-stopper flaw.
I recently succeeded in an 18-month-long effort to coordinate moving off of JIRA at work. We moved to clubhouse.io and it's been great. We were also looking closely at GitHub Issues + something like ZenHub for the missing project-management features — and that would have been great, too.
As soon as we moved off of JIRA, the amount of communication that started happening in the most relevant place — the bug tracker thread — increased by like... I don't know, something like 20x or 50x.
In our years on JIRA (the hosted plan) virtually nobody every discussed the details of the bugs on JIRA. We filed tickets there, and checked them off when done, and used it to review what had gotten done after each iteration — but nobody used it day-to-day. If discussion was to happen about a bug, it happened on Slack (which is horrible for ever finding it again, but it is fast).
The reason, I believe, is simply that JIRA is waaaayyy tooooo slow. Nobody wanted to use it day-to-day. You made a list of stories at the beginning of the iteration, and you checked it again at at the end of the iteration, to mark the ones you did done. I did this too.
Now, all the discussions directly about the bugs/features in the tracker happen directly on the tracker. Which, I recalled, was how it used to be several years ago before we moved to JIRA (from GitHub issues).
There is a lot of other stuff that is clunky about JIRA but I think we could have probably dealt with it, if it wasn't so slow that every interaction — even simple things like "find and open the bug I was working on most recently" or "see what is assigned to me this iteration" — felt like rm -rf node_modules && npm install and you somehow find yourself checking twitter before it even opens...
> I think you are right: the SLOWNESS is a killer. In fact, I think it is THE show-stopper flaw.
I think there are multiple clones of JIRA that literally just focus on the performance part. I think at this point it's unrealistic to expect major performance improvements in a timely fashion with JIRA.
If you want to try something in this space that's really snappy and responsive, checkout out Kitemaker (YC W21) [0]. I'm a founder and happy to chat with anyone that has any feedback/feature requests: hi@kitemaker.co
I was looking at the homepage reading text that explained the Picture Gallery ("when users can reuse images... more likely to...") and then the whole page refreshed with a completely different design. I was on the page for probably about 30 to 60 seconds.
Hi! That’s just a video playing on a loop taking you through the various screens. The “image gallery” is just showing our work item screen and then it jumps to some other screens. Thanks for the feedback though! It may be that this isn’t clear enough
Some constructive advice, your homepage is messy, the ratios and spacing are really off and makes it hard to digest and focus on your features. This will directly impact your actual product. Please don't be insulted , I've spend 20 years in design/branding and want you to succeed but this homepage isn't doing it, you need a brand book with some strict guidelines.
Sniff around a bit and you will find that what is even worse is why it is slow. Even the most simple operations call out to multiple sites. Quitting Jira requires more than three dozen sites to respond in order to actually work. Jira is an example of just how much the modern APIs are fun yay methodology sucks in actual practice.
It's funny because I have never seen the mentioned Firefox error on any Jira instance. The problem is mostly server-side: it's slow as hell! I'm personally using go-jira [0], but even that is slow: the problem as I said is not (only) the frontend, but the backend.
For most it seems "missing easy user interface" meaning there is an interface. And what they're wanting to do is part of a complex system therefore it's easy to understand why an easy interface isn't there, it's not a simple thing in the first place.
I've made the switch back to Redmine for my personal projects. By now it has full Markdown support for issues and wiki. Faster than Jira, and easier to use.
The new rich text editor still confuses me when it comes to newlines and quote blocks, in months since it's been gradually rolled out, I still can't remember if it's Shift+Enter or Enter to end the quote block and what kind of newlines work with quotes, if it's the selected text or a hidden paragraph that quotes are applied to and so on. Just let me edit the raw markdown as before or expose the underlying markup model when editing.
It has a markup language that accomplishes the same thing effectively but has a completely different syntax. This is super annoying when you are bouncing between pull requests in Atlassian Bitbucket and JIRA tickets. I have tried to create a heading with # only to end up with an <ol> at least a hundred times.
I imagine this is because JIRA (2002) predates Markdown (2004), or at least its mainstream popularity when GitHub adopted it.
Is that an excuse? If we have to make a fair comparison, HTML 5 is newer than both: is that a reason why one shouldn't not update the procuct?
As an additional side note, Jira updates are NOT free. You have to pay for a license fee every year to get the updates + support, but the product unfortunately remains kind of the same: super slow and with missing basic features (e.g: Markdown)
Trello is great for small projects and for personal lists and boards. I certainly wouldn't want to manage a complex project with a large software team from Trello.
Great article. As you are likely not native English, just as a tip you can actually simplify a lot of statements by removing the word ‘the’. As one example:
> When the users need the features from Jira and request them to configure the plugins for many application scenarios, which contribute to the overall complexity of UIs.
Becomes
> When users need features from Jira they request plugins, which contributes to UI complexity.
The first has incorrect use of the definite article (the users, the features), but to be honest it’s less about specific grammar syntax/rules and more about making writing easier to understand and more concise.
When you say “need the features”, that immediately makes me think “which features?” because it reads as if you are referring to some previously mentioned features.
It would be nice if any of the alternatives would be in line with Jira cost for a large enterprise. I can’t recommend this to anyone if our cost for issue tracking would quadruple overnight. Not to mention the migration costs...
I haven’t used any of those IDEs, but really, is Minecraft your example? A game with graphics so simple it should work on a Pentium 4 but requires a very beefy computer instead because it’s written in Java.
They rewrote it (not sure if in C++ or C#) for the windows store and it works perfectly on computers that can barely run the Java version.
You might not be familiar with how modern computer graphics work. Objects and textures are loaded into the GPU and managed with an API like OpenGL and shader programs written in languages such as GLSL. Java, in this case, serves as the engine that ties everything together and runs the core game logic.
Microsoft rewrote it because they're not a Java shop. Microsoft Games Studio not only uses C#, but their mission is to increase adoption throughout the industry as it feeds into Microsoft's ecosystem. Microsoft is about developer mindshare.
C# and Java are incredibly similar languages. They're almost indistinguishable. Pointing to C# as if it's substantially different from Java makes me think you've never used either.
I know exactly how it works. Java runs the game logic, and it does so so slowly that the game skips frames. If you’re claiming that the game logic and the code that runs the graphics are completely untied and run independently that’s wrong. The graphics code cannot paint another frame if it does not know where the objects are because the game logic, written in Java, is lagging behind.
C# is much faster than Java. Comparing them in real life performance situations (and not synthetic benchmarks) is almost a joke.
As opposed to C#, Java (alongside C/C++) is widely used in low-latency applications for high-frequency trading and HPC. Almost no one uses C# for high performance software.
Source: have been doing low-latency engineering for over 15 years.
It's quite apparent your experience with Java language/runtime is outdated and/or shallow. Maybe try reading up on it before making strongly opinionated comments like this?
Java minecraft did have issues. Professional game streamers were having issues due to the pause-the-world garbage collection.
I don't mean to over hype the issue, minecraft became on of the most successful games in history riding on Java. But the re-write to C++ was not just because Microsoft is a C++ shop. The C++ rewrite delivered measurable performance gains, making it much better on low end devices especially.
What is missing in my opinion? It is consistency between the atlassian products and/or adherence of industry “standards”
Confluence and JIRA use a different markup language for posts and none of them uses Markdown (in any flavour), which I would also use to write in-code documentation, like in a README file
The larger problem is that you cannot (AFAICT) edit the markup directly in Confluence or JIRA. I believe this used to be possible, but has been removed in the cloud version, and will die with the server version.
The markup you write gets turned into in-editor "objects" (for lack of a better term) as you type it out. This makes editing an incredibly painful experience. I am occasionally asked to document something on Confluence for a client and every time I feel like I am fighting a battle just to make a basic document.
For example, I strongly dislike their implementation of ordered list editing. They have tried to make the ordered list editing mode "smart", but it mostly just gets in the way, and is harder than writing it out manually. Additionally, you cannot embed other "objects" in an ordered list. If you attempt to insert a file object, or something similar, it splits your ordered list into two, with the second list starting its ordering at the beginning. This would seem to defeat the purpose of having a rich editor with embeddable objects, since I cannot use them if I am making an ordered list, such as set of "how to" steps.
I usually give up an hour in and switch to writing in my text editor instead, but even then Confluence will mangle my input in surprising ways when I paste it in. Rule #1 of making an editor in my book is that it has to be, at a minimum, as easy and useful as a basic text editor. Microsoft Word and LibreOffice, in spite of their faults, are quite good at this. I would guess most software shops could not implement a WYSIWYG editor at that level of polish, and they would be better off just providing editable markup.
The community/FOSS version mainly lacks the kind of reporting that jira has (the paid/Enterprise version has some basic reports, such as burndown charts). It's not perfect, but it's fast, simple and has an easy API and webhooks.
In the field I work in (CRM/webdev/consulting), JIRA used to be everywhere. Now all those shops have moved to Gitlab.
Phabricator is pretty good except it has pretty much zero CI integration. Kind of annoying.
You also have to pay to even comment on issues. I think that's a pretty great way of getting companies to pay for support but it's also annoying if you want to work with other freeloaders on issues.
Also it's written in PHP which means you aren't ever going to want to read or modify the code. No different to GitHub, or probably even GitLab but I can definitely imagine modifying Gogs/Gitea.
OP mentioned "non-cloud", so I assumed they wanted it self-hosted. You don't have to pay anyone for commenting in Phabricator when you're self-hosting it.
Err, I think you misunderstood. You have to pay to comment on issues about Phabricator. On their official bug tracker. Like, if you find a bug, or a solution for a known bug, you can't tell them without paying.
Doesn't have anything to do with cloud hosting our self-hosting.
I would like to try a more visual and spatial approach to project planning software. Most of the major solutions, like JIRA, are more like interacting with a RDBDMS through a UI. (Although I admit JQL is very useful, so I don't want to throw that away entirely.)
Here's another product I found recently that allows users to draw out their project plan, and have it turned into management-friendly reports automatically:
Yeah, that's right. Before we escaped JIRA, my team would often do that, too, even for regular sprints. Make a checklist in a collaborative editor tool like Google Docs or Notion, and just do the day to day work there.
Then somebody would have to copy it to JIRA so the work would get tracked.
But very little of the planning, work breakdown, day to day updates, or discussion actually happened on JIRA because it was literally like a hundred times slower (with the convoluted UI compounded by incessant "wait... is it even loading?... oh, OK, it finally loaded" at each step of the way).
I manage a team of devs but my clients love to use Jira. So I typically use something like Board Genius to sync between Github issues and Jira. This also let's me keep my devs out of Jira so they can just focus on actual work and not worry all the discussions that happen in Jira between my clients and project management.
My workplace uses Monday.com and some folks like it but I personally find it clunky and underwhelming. In my opinion, it tries to do too many things and doesn't do any one thing particularly well. If you are just wanting a kanban board, you are better off sticking with Trello. If you are wanting bug tracking, feature requests, etc, you are better off with things like GitHub Issues and other specialized tools. YMMV of course.
If you want a normal software engineering process, where you want to check in with your colleagues after 1-2 weeks on the progress etc. it's not going to happen. The normal concepts of sw. eng. aren't in Monday. It's more like a clunky to-do for non-techy people. Simple things are extremely hard in Monday. For example seeing all the tasks from all projects assigned to you -> hard. I'd rather stick with GitHub/JIRA
Co-founder of Zepel.io [0] here. Thank you for giving Zepel a shoutout in the Jira alternatives section of the landing page. Would love to get feedback from the community regarding our App.
As others have said here, the main problem is just that it’s really slow.
Apart from that, I think the only item on this list that really puts its finger on a key problem is #1, inability to manage dependencies between projects. You try to work around Jira’s problems by splitting the work into separate projects, but then they’re almost completely siloed, and Jira doesn’t help you connect them up.
The other problem I have with it is that it’s too fussy about the hierarchy of issue types. You have sub-tasks, stories, epics, and initiatives. Each one has special semantics and certain supported operations. If you started defining some work at the wrong level of granularity, too bad, because switching everything to a different tier is complicated and massively slow (literally a multi-page wizard, a progress bar and a confirmation screen when making a “mass edit”, i.e. a trivial edit to two or more issues).
Hierarchical issues is a good idea, because that’s a very natural way to break down and refine work estimates, but it should just be an arbitrary tree, not a strict and limited set of tiers.
Too much complexity, not enough generality. (And not enough speed.)
Looks like the complexity is a consequence of its generality. Much like Word it serves so many different cases that many people end up using only subsets of features. Yet the code must support all possible cases.
Competitors too will likely drift into the same quagmire unless they're unashamedly opinionated.
What’s killing Jira and the other tools in Atlassian’s suite for us is them killing off the on-premises server products. Some of the companies I work with have defense security requirements and just can’t use it in the cloud (unless they followed a lot of stringent requirements that are not on their roadmap, and I assume they would charge a lot more for that).
The cloud versions are also far more expensive and while I’ve always liked the software in general (apart from it being dog slow), a big draw was how cost effective the server products were. So I’m about to start evaluating GitLab, YouTrack, etc. to start moving these companies off Jira, Bitbucket and Confluence in the next year or so.
Let me walk that back a bit - I have no experience of the cloudy version which unanimously seems to be derided as slow. Atlassian - you might wanna fix that. Like yesterday.
But server (& datacentre) editions - best in class for my money.
In terms of features, the Rally / Agile Central (did that just re-brand again?) rollup roadmap view is the only missing feature i crave. Atlassian should just merge the portfolio product into jira.
As for everything else - speed is absolutely fine if hosted well.
I tend to council people to try to use less features not more - often jira features are used to attempt to work around challenges that are best addressed by talking to people instead.
It integrates with everything, no matter what source control, build pipeline or IDE tooling is in use, there’s integration with jira.
The REST API is absolutely usable (mostly for workflow automation and analysis purposes).
Server starts heading towards EOL in February 2021 unfortunately. They’ve said that they will support it for three years after that date but you’ll only be able to renew licences (not buy new ones), the pricing is going up and you can’t change tiers.
Datacenter is still supported but I think they’re jacking up the price of that too.
> But server (& datacentre) editions - best in class for my money.
Well, performance wise they're as bad as the cloud version. Even when you give the instance a lot of CPU / Memory. The problem is not only the Spring application itself, the DB seems to be missing some indexes or the queries are super badly optimized. God only knows what's really happening under the hood, but my self hosted Jira instance isn't any better (I've tried them both, cloud and on-prem).
No, cloud is slow and has been slow for years. On-premise is quick enough, but only if you ensure you have no slow bottle-necks (DB, spinning disks, heavily shared virtualization).
Jira doesn't use Spring--you may be confusing it with Confluence which does.
The Structure add-on does a great job of rolling up. The Portfolio product was way too difficult to use--they rebuilt it so maybe things might be different now.
I agree there are/were the best in class for the money--but with Atlassian throwing all their resources into cloud now, raising the prices on on-premise, I worry for its future. I suspect we're only going to get timid features--nothing bold that might improve the UX/UI, or anything that will break the migration path to Cloud, which is their end-goal.
If I was starting out again, I'd probably examine YouTrack where there's at least some assurance they are investing in its future.
I kinda agree. I am fine with using Jira. Now when I know the workflow I would not bother with another tool for the same things that we are forced to use Jira for.
That said I would rather to "issue tracking and planning" in a checked in text file and maybe have some bug tracking tool rather then working with Jira.
I also have a low opinion of jira, but I don't agree with this list at all. I refuse to use jira because it is over-engineered and is generally used as a crowbar to foist non-technical influence upon technical teams, at least in the sites I've seen. Jira is in my opinion a fatal source of friction to the production of reasonable software.
That's not what I'm getting from this list.
Worse, almost all of the points (all maybe? Haven't checked) declare that something is "missing".
Please god no.
Jira isn't missing more features. When you make a list like this, some product manager at Atlassian is going to turn it into a bunch of scrum epics and jira will be even worse in 12 months.
Yup. Jira seems to be a tool around creating process. Management loves that shit because they believe if they put enough checkboxes and fields into a ticket eventually things will be good (or something). It drives me nuts to no end how stupidly complex people want to make the development process.
What should, IMO, have almost no interaction with the ticketing system, ends up being a chore for development which adds 0 value to the final product.
If anything, Jira tries to do too much. It is like Excel, tries to fit every need in its domain. That it still has limits does not mean it sucks. Quite tiring when people reach for that word for nitpicks and personal opinions.
529 comments
[ 0.22 ms ] story [ 374 ms ] threadThe only bullet point I do agree with, is that Jira is missing an easy (or IMHO a minimalist / simple) interface. The tool is a mess, and the people using it are probably also using it in the wrong way, making the whole situation way worse.
Don't get me wrong, I like Jira to some extent, but its performance and messy UI are really making my life harder rather than simpler.
They also hid bugs from being counted in reports by marking them as stories but that's another conversation
These tools market themselves as creating transparency but they are in practice often used to hide or excuse things.
I suppose that is why they all take the tragic path from being heralded as new and fresh in the beginning to absolutely being hated in the end.
Goes for methodologies as well (I remember the agile manifesto ... look at where we are now).
There's no way to cater every possible workflow without feature bloat. However, the space is ripe for simpler, more accessible solutions that supports work across multiple projects.
Atlassian products are all like this, and each one is a horrible trash fire.
Confluence is the worst wiki product I've ever used. Gross syntax, slow, clunky plugin system, awkward permissions. Install MediaWiki and be done with it.
Bitbucket is slow and feels like it's from the 90s. It doesn't seem to do multithreading well - sometimes force pushed branch rebases pick up a new, unrelated author. A bug that hasn't been fixed in years, and only one of many. How many times will merging not actually work?
Jira is a special kind of hell. It's like every team needs their own embedded Atlassian solutions engineer to get it into a usable state. Not an EM or PM, but a specially dedicated role just to deal with Jira nonsense. Why even bother?
Atlassian is like Salesforce for business process, but if it were written by all-remote temp contractors churning through Jira tickets themselves.
The horrible response times and user hostile UI is shocking, especially if you look at their competitors. It gets in your way all the fucking time. But because they’re embedded in enterprises they don’t need to change much. It’s a big mess.
BUT... the direct integration with the ticket system - highlight a list, and make linked tickets (which keep the status info 'live' in the confluence doc) is pretty powerful, and allows a lot of people to keep their business docs, and get status updates without ever having to learn other parts of the system.
[0]: https://linear.app/
The complexity comes from the demand of Jira's paying professional customers.
Eventually Jira will grow too complex and bane the way for simpler tools probably of a different paradigm. Much like how Jira replaced things like HP Quality Center or the tools of Rational Unified Process.
I recently succeeded in an 18-month-long effort to coordinate moving off of JIRA at work. We moved to clubhouse.io and it's been great. We were also looking closely at GitHub Issues + something like ZenHub for the missing project-management features — and that would have been great, too.
As soon as we moved off of JIRA, the amount of communication that started happening in the most relevant place — the bug tracker thread — increased by like... I don't know, something like 20x or 50x.
In our years on JIRA (the hosted plan) virtually nobody every discussed the details of the bugs on JIRA. We filed tickets there, and checked them off when done, and used it to review what had gotten done after each iteration — but nobody used it day-to-day. If discussion was to happen about a bug, it happened on Slack (which is horrible for ever finding it again, but it is fast).
The reason, I believe, is simply that JIRA is waaaayyy tooooo slow. Nobody wanted to use it day-to-day. You made a list of stories at the beginning of the iteration, and you checked it again at at the end of the iteration, to mark the ones you did done. I did this too.
Now, all the discussions directly about the bugs/features in the tracker happen directly on the tracker. Which, I recalled, was how it used to be several years ago before we moved to JIRA (from GitHub issues).
There is a lot of other stuff that is clunky about JIRA but I think we could have probably dealt with it, if it wasn't so slow that every interaction — even simple things like "find and open the bug I was working on most recently" or "see what is assigned to me this iteration" — felt like rm -rf node_modules && npm install and you somehow find yourself checking twitter before it even opens...
Speed matters!
I think there are multiple clones of JIRA that literally just focus on the performance part. I think at this point it's unrealistic to expect major performance improvements in a timely fashion with JIRA.
0: https://kitemaker.co
Firefox 84 on Kubuntu Linux, no relevant addons.
It says "missing" but it mostly talks about those existing features lacking any polish.
> and that it's super slow!
It is and it really sucks or to quote Firefox "A web page is slowing down your browser, what do you want to do, wait or stop it".
[0]: https://github.com/go-jira/jira
or
https://trac.edgewall.org/
- YouTrack: https://www.jetbrains.com/youtrack/
I imagine this is because JIRA (2002) predates Markdown (2004), or at least its mainstream popularity when GitHub adopted it.
As an additional side note, Jira updates are NOT free. You have to pay for a license fee every year to get the updates + support, but the product unfortunately remains kind of the same: super slow and with missing basic features (e.g: Markdown)
- Whatever GitHub/GitLab provides if I need the most basic level of tracking.
- Trello for a more polished experience, a few additional ticket/board related features and easier involvement of non-developers.
- Linear if there's a bigger team that needs more than a bunch of boards (backlogs, roadmaps, projects, teams etc.).
I wholeheartedly agree with your suggestions. Trello won't scale, but it's good for n < 3 users, especially if working with non-engineers / non-PMs.
On android, returning from a screen unlock deleted all my unsaved text.
https://joearms.github.io/published/2014-06-25-minimal-viabl...
In use from 1985 to 2014 and maybe longer.
> When the users need the features from Jira and request them to configure the plugins for many application scenarios, which contribute to the overall complexity of UIs.
Becomes
> When users need features from Jira they request plugins, which contributes to UI complexity.
Not a criticism - just thought this might help!
When you say “need the features”, that immediately makes me think “which features?” because it reads as if you are referring to some previously mentioned features.
I found it interesting that the first bullet point (at least when I loaded the site) was:
That indeed is why Jira sucks, in my opinion (and apparently many other people, judging by comments in this thread).Minecraft Java Edition is fantastic.
Java is just a tool. Atlassian isn't a good workshop.
They rewrote it (not sure if in C++ or C#) for the windows store and it works perfectly on computers that can barely run the Java version.
Microsoft rewrote it because they're not a Java shop. Microsoft Games Studio not only uses C#, but their mission is to increase adoption throughout the industry as it feeds into Microsoft's ecosystem. Microsoft is about developer mindshare.
C# and Java are incredibly similar languages. They're almost indistinguishable. Pointing to C# as if it's substantially different from Java makes me think you've never used either.
C# is much faster than Java. Comparing them in real life performance situations (and not synthetic benchmarks) is almost a joke.
Source: have been doing low-latency engineering for over 15 years.
It's quite apparent your experience with Java language/runtime is outdated and/or shallow. Maybe try reading up on it before making strongly opinionated comments like this?
I don't mean to over hype the issue, minecraft became on of the most successful games in history riding on Java. But the re-write to C++ was not just because Microsoft is a C++ shop. The C++ rewrite delivered measurable performance gains, making it much better on low end devices especially.
Confluence and JIRA use a different markup language for posts and none of them uses Markdown (in any flavour), which I would also use to write in-code documentation, like in a README file
The markup you write gets turned into in-editor "objects" (for lack of a better term) as you type it out. This makes editing an incredibly painful experience. I am occasionally asked to document something on Confluence for a client and every time I feel like I am fighting a battle just to make a basic document.
For example, I strongly dislike their implementation of ordered list editing. They have tried to make the ordered list editing mode "smart", but it mostly just gets in the way, and is harder than writing it out manually. Additionally, you cannot embed other "objects" in an ordered list. If you attempt to insert a file object, or something similar, it splits your ordered list into two, with the second list starting its ordering at the beginning. This would seem to defeat the purpose of having a rich editor with embeddable objects, since I cannot use them if I am making an ordered list, such as set of "how to" steps.
I usually give up an hour in and switch to writing in my text editor instead, but even then Confluence will mangle my input in surprising ways when I paste it in. Rule #1 of making an editor in my book is that it has to be, at a minimum, as easy and useful as a basic text editor. Microsoft Word and LibreOffice, in spite of their faults, are quite good at this. I would guess most software shops could not implement a WYSIWYG editor at that level of polish, and they would be better off just providing editable markup.
The community/FOSS version mainly lacks the kind of reporting that jira has (the paid/Enterprise version has some basic reports, such as burndown charts). It's not perfect, but it's fast, simple and has an easy API and webhooks.
In the field I work in (CRM/webdev/consulting), JIRA used to be everywhere. Now all those shops have moved to Gitlab.
You also have to pay to even comment on issues. I think that's a pretty great way of getting companies to pay for support but it's also annoying if you want to work with other freeloaders on issues.
Also it's written in PHP which means you aren't ever going to want to read or modify the code. No different to GitHub, or probably even GitLab but I can definitely imagine modifying Gogs/Gitea.
OP mentioned "non-cloud", so I assumed they wanted it self-hosted. You don't have to pay anyone for commenting in Phabricator when you're self-hosting it.
Doesn't have anything to do with cloud hosting our self-hosting.
https://discourse.phabricator-community.org/
[1] https://spartez-software.com/products/whiteboards-for-jira
Here's another product I found recently that allows users to draw out their project plan, and have it turned into management-friendly reports automatically:
https://www.gameplan.global/
No comment on whether or not this is actually good, since I have never tried it, but it presents a refreshing concept.
2. It is dog slow.
3. It is dog slow.
...
When my team had an important as-quickly-as-possible sprint, we dropped Jira and switched to Google Docs as it was much easier and faster than Jira.
My conclusion was that Jira doesn't do much for developers that can't be done in some Google Docs/Sheets.
Then somebody would have to copy it to JIRA so the work would get tracked.
But very little of the planning, work breakdown, day to day updates, or discussion actually happened on JIRA because it was literally like a hundred times slower (with the convoluted UI compounded by incessant "wait... is it even loading?... oh, OK, it finally loaded" at each step of the way).
We had a product manager who used to write tickets in OneNote and then copy and paste into JIRA when things were quieter in the evening.
We build it so that almost everything you do in the app instant (<100ms)
[0]:https://zepel.io/
Apart from that, I think the only item on this list that really puts its finger on a key problem is #1, inability to manage dependencies between projects. You try to work around Jira’s problems by splitting the work into separate projects, but then they’re almost completely siloed, and Jira doesn’t help you connect them up.
The other problem I have with it is that it’s too fussy about the hierarchy of issue types. You have sub-tasks, stories, epics, and initiatives. Each one has special semantics and certain supported operations. If you started defining some work at the wrong level of granularity, too bad, because switching everything to a different tier is complicated and massively slow (literally a multi-page wizard, a progress bar and a confirmation screen when making a “mass edit”, i.e. a trivial edit to two or more issues).
Hierarchical issues is a good idea, because that’s a very natural way to break down and refine work estimates, but it should just be an arbitrary tree, not a strict and limited set of tiers.
Too much complexity, not enough generality. (And not enough speed.)
Competitors too will likely drift into the same quagmire unless they're unashamedly opinionated.
* Google Analytics
* Google Search Console
* Jira
So basically I wait and wait and wait to manage projects to make websites fast.
The cloud versions are also far more expensive and while I’ve always liked the software in general (apart from it being dog slow), a big draw was how cost effective the server products were. So I’m about to start evaluating GitLab, YouTrack, etc. to start moving these companies off Jira, Bitbucket and Confluence in the next year or so.
Let me walk that back a bit - I have no experience of the cloudy version which unanimously seems to be derided as slow. Atlassian - you might wanna fix that. Like yesterday.
But server (& datacentre) editions - best in class for my money.
In terms of features, the Rally / Agile Central (did that just re-brand again?) rollup roadmap view is the only missing feature i crave. Atlassian should just merge the portfolio product into jira.
As for everything else - speed is absolutely fine if hosted well.
I tend to council people to try to use less features not more - often jira features are used to attempt to work around challenges that are best addressed by talking to people instead.
It integrates with everything, no matter what source control, build pipeline or IDE tooling is in use, there’s integration with jira.
The REST API is absolutely usable (mostly for workflow automation and analysis purposes).
Datacenter is still supported but I think they’re jacking up the price of that too.
If Jira takes a performane hit by lowering the bar for the non-techs, fine.
Just use the tool.
Well, performance wise they're as bad as the cloud version. Even when you give the instance a lot of CPU / Memory. The problem is not only the Spring application itself, the DB seems to be missing some indexes or the queries are super badly optimized. God only knows what's really happening under the hood, but my self hosted Jira instance isn't any better (I've tried them both, cloud and on-prem).
Jira doesn't use Spring--you may be confusing it with Confluence which does.
I agree there are/were the best in class for the money--but with Atlassian throwing all their resources into cloud now, raising the prices on on-premise, I worry for its future. I suspect we're only going to get timid features--nothing bold that might improve the UX/UI, or anything that will break the migration path to Cloud, which is their end-goal.
If I was starting out again, I'd probably examine YouTrack where there's at least some assurance they are investing in its future.
That said I would rather to "issue tracking and planning" in a checked in text file and maybe have some bug tracking tool rather then working with Jira.
That's not what I'm getting from this list.
Worse, almost all of the points (all maybe? Haven't checked) declare that something is "missing".
Please god no.
Jira isn't missing more features. When you make a list like this, some product manager at Atlassian is going to turn it into a bunch of scrum epics and jira will be even worse in 12 months.
What should, IMO, have almost no interaction with the ticketing system, ends up being a chore for development which adds 0 value to the final product.