Kerning and auto-widening

[Uffe]

Hi all,


If an element name is wider than what can fit into the name text box (and it doesn't have hypens or spaces where newlines can be inserted, and the name isn't too ludicrously long), EA widens the element to accomodate it.

Problem is, it doesn't widen the element enough, so the letters in the name end up scrunched together. Which you can fix quite simply in the GUI.
Which makes you wonder why EA bothers to widen the element in the first place.

This is possibly less of an issue when modelling in English than in Swedish, German, Finnish, etc. But still.

Could the auto-widening be fixed?
Or should we perhaps aim higher and think about introducing auto-hyphenation instead?

Discuss.


/Uffe

[qwerty]

I remember this being discussed in the (distant) past. I would not expect some outcome from the Sparxians. There are so much silly things about rendering that should be fixed. But they are probably just designed the wrong way. Haven't seen some real fix for wrong design in EA (except maybe in the pre V6 versions).

q.

[MichaelJ]

...Haven't seen some real fix for wrong design in EA (except maybe in the pre V6 versions).

Based on your experience, I think this may imply there is no more quality development/management team at Sparx whose (personal) goal is to improve the product -- the "good guys" left many years ago. So now, paying customers are stuck with an aged, lethargic (possibly cheap-outsourced) dev. team who shy away from actually fixing bugs and improving the user experience. Considering they boast 850K registered customers (let's assume each customer pays the least subscription amount of $299 USD), it's baffling to see the EA product still languishing under the weight of poor design decisions and loads of bugs (such as the one raised by /Uffe).

...Could the auto-widening be fixed?....

You could always log this as a bug and then hope that the current support team care enough about your concern to actually fix it in an upcoming release. My experience has shown, that no matter how deep the bug analysis provided to the current dev/support team in helping them to RESOLVE and FIX a bug, it's always the same outcome: never addressed.

[qwerty]

Yeah. It's that marketing team from Sirius Cybernetics Corp. which started ruling Sparx a long time ago. Ever since I see just feature creep and no bug fixing. When does this revolution come?

q.

[MichaelJ]

Yeah. It's that marketing team from Sirius Cybernetics Corp. which started ruling Sparx a long time ago. Ever since I see just feature creep and no bug fixing. When does this revolution come?

Wow, loved the link thank you! The first few opening sentences reflect the truth we have to face daily in EA. The statement "...not known for the quality of their products, and almost all of their known inventions are faulty..." is pertinent.

As for "Revolution"? Major external big-name reviewers such as Gartner and their "Magic Quadrant" provide us with an unbiased business-case assessment of vendors at play within a given market segment. I think the only reason (or opinion) EA survives to this day, is their relatively low cost to entry. Also, the general lack of a viable and cheap competitor with similar or better functionality, user experience and feedback system. I do wonder, as more customers use the product, and note it's awful and inconsistent user experience, surely they must demand better software from EA?

In demanding better software from a vendor, customers help by logging bugs and features. Unfortunately, in EA's scenario, bugs are never fixed, and feature requests might as well not exist. The team seem to operate with arrogance as though separate from paying customers, such as you or me. No matter how deeply you provide insight into the nature of EA's software bugs (and potential resolution), the development team know better. And release-after-release, bugs remain and UX remains stagnant and difficult.

"Revolution" will come when customers reach the "point of no return" and decide that the effort made to support the product far outweighs the "rewards". Put financially, customers stop paying for a license to use a product that no longer serves their modelling needs in the way they require.
 
 

[Geert Bellekens]

I don't completely agree with the sentiment here.

I've had a number of bugs (and a few feature requests) fixed for me, even in the last half year.
I'm not saying Sparx can't (and shouldn't) do better, but it's not all gloom and doom either.

And to put things in perspective; I have experience with some of the other tools in Gartners magic quadrant report, and I have serious doubts about their assessments.

MEGA for example always scores somewhere in the top right quadrant, and, based on the Gartner reports seems like a viable competitor for EA.
Now MEGA is the absolute worst of the worst.
- Working with this product feels like you've stepped in a time-machine back to 1997.
- It's "database" can only hold a few thousand elements before it becomes unworkable.
- It regularly has data consistency issues making it loose data
- It can't even export to XMI properly without running into exception after exception
- It's terribly slow (waiting 10 minutes for your model to open is considered "normal")
- ....

Sparx Systems and Enterprise Architect definitely have much room for improvement, but I'm not convinced other tools and vendors are any better.

I still consider EA to be the best general purpose modelling tool out there, not even considering the price;
Although probably: "In the land of the blind, one-eye is king."

Geert

[Eve]

This is an issue that comes up periodically, and it's not one that is easily solved.

The latest round of issues are as a result of the changes to scale up all diagrams a little by default. You can prevent most of the issues by setting that back to no scaling. That will likely result in fonts uncomfortably small for most users, you can set the model font to be a little larger, but you will need to check diagram layouts as a result.

For better or worse, in this context definitely worse, EA still relies on Windows font rendering. I could rant endlessly on that, but suffice to say, it is those APIs that are the limiting factor here and we can't achieve a perfect result without a completely different renderer. Part of the problem with that is that it will almost certainly change the base result, likely making every piece of text wider. That means all elements would increase in size, for all users.

[Uffe]

Not an easy one I agree, although I do feel that of all the GUI stuff in EA, diagram layout is the one that users spend the most time on for the least productive gain. If I had a vote, I'd cast it on this area to be the top priority for improvements. Much as I'd love to see better APIs and various other bits and bobs, diagram layout is where users in general can benefit the most.

What about auto-hyphenation? Has that been considered?

/U

[MichaelJ]

...I'd cast it on this area to be the top priority for improvements....

Absolutely. Even for just one release, it will be much appreciated that the EA team actively worked to address lingering and obvious major UI inconsistencies (scrollbars missing in resizable windows / tree views, lack of sorted entries, odd data entry order, inconsistent drag-and-drop functionality from the Traceability window).

... EA still relies on Windows font rendering... those APIs that are the limiting factor here and we can't achieve a perfect result without a completely different renderer...

GDI+ APIs are indeed dismal for UI font rendering/kerning/diagramming. Modern browsers can render 3D elements, shapes, perform complex transformations/effects etc. (think, CSS styling) much better than GDI+ APIs. The existing rendering engine in EA is indeed problematic and does not fit well with business users who have come to expect "Visio-style" quality. Since there is no use in trying to save a dead horse, one supposes the best approach is to engineer an HTML-based / DirectX-based that can be leveraged within EA and ultimately allow the aged horse to become one with the earth in due time.

For example, to observe rendering performance, please see:

• A JavaScript open source set of diagrams types: https://observablehq.com/@d3/gallery
• Database diagramming: https://dbdiagram.io/d
• General diagramming: https://app.creately.com/

...I'm not saying Sparx can't (and shouldn't) do better, but it's not all gloom and doom either...

Of course, hope is not yet lost! EA is not bad, I agree with you. Each release brings 00's more bugs. Isn't the point of "QA" to reduce bugs and improve a product? You, I and other customers act as external QA. We aren't paid for our time to log bugs, we do so because we hope that EA is not all doom and gloom 

It's good to hear some of your bugs have been addressed and you raise a good point concerning the MEGA product: just because Gartner lists a vendor as the top 5 vendors, does not necessarily reflect the the reality of that vendor's product.

[skiwi]

I don't completely agree with the sentiment here.

+1

[skiwi]

If I had a vote, I'd cast it on this area to be the top priority for improvements.

+ 1 and more generally for diagram improvements (see many previous threads)