Forecasting shows a massive gap in Visual Studio Team Services - azure-devops

I'm having really odd problems with my VSTS Forecasting tool - It doesn't seem to matter what I set the Forecast velocity to, it leaves a massive gap of around 20 stories before it starts doing anything. Has anyone else had an issue like this and can point me at what may be causing a problem?
I'm sorry there's not much more to go on but I can't offer much more without providing my product backlog and I can't do that.
I've tried changing the settings to only show Backlog Item level and below.
Here's a screen grab - I've had to do it in 2 parts because I actually have to scroll down to see any of the forecast lines...
Moving the items with no estimation value to the bottom of the backlog made no difference either (although I would expect a fair few more break lines before hitting that point)

The problem here is that (in most cases) I was using the 'Story Points' rather than the 'Effort' field (because that's what they are!). I fixed the issue by replicating all of my story points data into the 'Effort' field. The Story Pints field doesn't actually seem to do very much.
This article: https://www.visualstudio.com/en-us/docs/work/guidance/agile-process-workflow seems to disagree with this theory however.

Related

Recreate GitHub Network Graph in Java

I'm currently trying to recreate the GitHub Network Graph in Java as a school project. I'm using Kohsuke's Github-API to get all the data (repo, commits, etc...) and I actually got a prototype to work with some repositories out there, even though it's still not quite right.
My problem is getting the correct vertical spacing in the network graph.. My prototype is currently spacing the commits via some rules to detect merges/forks/branches but it's not working well with bigger and more complex repositories.
Here I found someone who recreated the old flash version of the network graph in JavaScript. But he is using a variable called "space" to determine the vertical spacing.
When I looked into the .js that creates the current Network Graphs I found that they also use a variable called space and spaceMap to get the spacing right.
It seems that without this variable it's not quite possible to make it look like the original Network Graph, but I didn't find the variable in Kohsuke's GitHub-API nor in GitHub's original API... But then again, I'm quite new to this GitHub-Story so maybe I'm just looking in the wrong places :/
If anyone tried to recreate the graph or knows how to find the spacing data I'd appreciate some hints.
Thanks

VB6 How to disable weekend days from Calendar control

I'm almost finishing my VB6 app and I'm trying to add some visual enhancements to controls.
My app allows to assign medical shifts to patients for which I use a Calendar Control. Now, I'm trying to find a way to disable Saturday and Sundays from being selected. Besides, it would be great if I could change the color of days if there are available shifts on them (that will be based on a sql query).
Does anyone know if this is feasible? I have looked at so many web pages for help but no success.
Thanks for your help
Gerardo
Absolutely not. The control is not even a standard Windows control, which means you're limited by the original implementation of it when Microsoft made it for VB6 back many years ago.
The easiest solution is to produce an error, or cause the cursor to switch to a different, nearby date.
The best solution is to make your own control, which would not be such a bad idea if you want anymore flexibility.
You could redesign your form to have a grid that shows appointment times, by staff. Just shade the squares according to availability.
Use the Weekday(date) or WeekdayName(date) function to see if the user entered a date that corresponds to a weekend day.

JCAPS little boxes of hell

Has anyone out there in SO land had to deal with this...when you go to build a new deployment profile in JCAPS, you have quite a few little boxes that you have to minimize before you can do the actual mapping (black boxes added to hide company specific info):
We've tried autohotkey and are looking into Sikuli - does anyone know of a good way to get around this silly UI design flaw?
You can build multiple environments that point to the same logical host. Just separate the resources into logical groups to make them more manageable.
I wrote a Netbeans plugin to take care of this. I can't publish the full source, but here is the main idea.
I've found that if you click on the "third" button in the deployment wizard that your logical hosts appear on the lower right. Other than that, the latest EAs you created are on the upper left.

Learning Eclipse MDT-UML2 Sequence Diagrams

Are there any tutorials or documentation files for drawing Sequence Diagrams with Eclipse MDT-UML2?
I can't put a Lifeline on the page. It took me 10 minutes to figure out that I first have to make a big Iteraction, and then I can put Lifelines into it.
I can't figure out how to make the Participant box wider in the lifeline. Any name more than 4 characters is too wide for the box. When I do the obvious thing of clicking and dragging the handles, the box springs back to its original size.
I can add a Message by clicking "Message" in the pallette and click-dragging an arrow from line Lifeline to another. Now I have an arrow labeled "1: *". I've tried every way I can think of to change that name: double-clicking on it, shift-clicking, ctrl-clicking, changing Name in the properties.
Can anyone help?
As I understand it MDT-UML2 is in the very early stages, it may not work as it should always. Additionally, the sequence diagram is a newer one, see below. There were no tutorial I could find, but I did link to samples below. I also found a comment in the eclipse bug tracker for this feature:
Here are the most important issues from my side: moving of lifelines, create messages,
replies are causing strange behaviour (some kind of auto alignment). Moreover it
should be possible to give the messages names.
They say it is resolved, but you be the judge.
From the wiki (Note:What you want is UML2 Tools):
http://wiki.eclipse.org/MDT-UML2Tools
Interaction diagrams (in progress):
Sequence (new in Galileo)
This may not help, but here is a link to the samples: http://wiki.eclipse.org/images/3/3a/UML2Tools_samples.psf
The tutorial reference does not seem to cover the new stuff, just the class diagram.
http://www.vogella.de/articles/UML/article.html
Search stackoverflow for other open source tools to use. The MDT-UML2 is coming along and can be used for some stuff, especially simple DSMs/DSLs/UML Profiles, but for drawing easily there are a host of good tools. An ok thread: What's the best UML diagramming tool? or
https://stackoverflow.com/questions/390438/good-free-uml-tool-for-java-eclipse

TODO/FIXME plugin for Eclipse

In my project there are large no. of FIXME / TODO which are to addressed at some point of time. Actually there about 480 which can be seen from 'TASKS' list but not organised.
I googled and found the Task Tag Decorator plugin.
But unfortunately this is not working.
Can anyone suggest a plugin for FIXME/TODO
apart from this.
I would also wanted to hear from all how these situations are usually managed
What you don't want to hear is that how these situations are usually managed is by not letting them grow so big. But I'm afraid that is the case.
The Pragmatic Programmers advise us Don't live with broken windows. The point being, that if we leave something broken instead of fixing it then other things will be left and before we know it we have 480 items on our TODO list. Plus, there is a danger that some part of our application will come to rely on the "broken" behaviour, so when we address the TODO item we also have fix that as well.
Not everybody can live up to the Pragmatic Programmers' high standards. An alternative approach is to have a list of stuff which needs to be worked on (sometimes known as the Kaizen list). People who are blocked on their assigned work can pick up one of those tasks.
As for your current situation....
I have a rule of thumb which states that nothing can be done in less than half-a-day: not once you include source control, documentation, discussing the change with Bob, etc. Of course, my rule of thumb doesn't apply to truly trivial tasks, but if these tasks were truly trivial they would have been fixed on the spot, not marked as TODO, right?
So you're looking down the barrel of 240 days of effort. If lots of those tasks can be combined into a single fix then you can reduce the per task overhead. But first you've got a chunk of work just to sift through the tasks, categorising and prioritising them. This is why thay call it "technical debt": the longer we leave it the more it costs to fix, and it has the compound interest rate of the average doorstep loanshark.
Unless you have a very understanding project manager/paying customer I think you will have to accept that you aren't going to be able to clear all these items. So you need a brief triaging exercise: assign each TODO into one of three categories:
Stuff that is intolerable and needs
to be fixed right now
Stuff that ought to be fixed as and
when there is an opportunity
Stuff that you're just going to have
to live with
Good luck!
Just wanted to chime in and say that I was able to get the Task Tags Decorator working in eclipse 3.5.0. Here's how.
Install from the update site : Task Tags Decorator Update Site
In preferences: General->Appearance->Label Decorators->Task Tag Decorators
Set up your decorators there.
Then go to: General->Appearance->Label Decorators and check the "Task Tag Decorator" box and Apply
I found this very frustrating to figure out and I hope it helps others. I really like the plugin and have found it helpful for me to visually keep track of TODOs. I think it also helps me keep the TODO list short since it's not out-of-sight-out-of-mind.
In what way do you want them to be organised? More specifically what doesn't the Task View do that you want to do?
For information you can do the following with the standard tasks view:
Click each column header in the task view to sort by that column ascending/descending.
Select the View Menu (downward triangle in top right of view) and
Group tasks by type
Show only a particular task type
Select Configure Contents... then :
Select a Scope to restrict the tasks shown to your desired scope (I find On selected element and its children particularly handy).
Filter tasks by description text (using contains or does not contain)
Filter based on Priority and/or Completed status and Task Type
There are also a few other options you may find useful if you dig through the View Menu.
You might want to look into Mylyn.
It might also be worth considering what it means to use one of these tags.
The trouble is that if you have almost five hundred of these things it doesn't seem likely that you are going to be able to delete them as 'stuff to do' at any point soon. Thus, the impact of using the tag and the process implied therein is diminished, and you end up with the problem you are citing, that the data set has become unmanageable.
The point is that the code is the truth, not the comments, whether they are marked up with task tags or not. You have to have something measurable by which you improve the code.
As an example, when I do code reviews I use these tags in a first pass before the code review pairing begins. The aim is that by the time the code review is complete the tags do not exist.