A new company, another set of random values for serialVersionUID.
Not sure why most programmers have the urge to put in some huge, ‘random’ number here. The objective is to match versions during serialization. Can’t imagine using a random number is a good idea. Even better, the numbers entered are not random.
Sorry, this is just a pet peeve. There is much written up on using this correctly if you are truly serializing objects. I think I have seen this at every company I ever worked.
Would you work where the programmers had to pass a personality test?
Total Voters: 4
Guess this is a little to close to the PC line for people to express their opinion. Too bad.
I have grown to really hate eclipse. Seems like every keystroke now starts an “update maven dependencies” task.
Have tried several of the options that supposedly stop this to no avail.
Doing most actual builds from the command line using mvn commands.
Starting to use notepad+ for editing files lest I start another update process.
Even worse, in many cases when this is happening I am changing very few lines of code.
Last bit was doing a search, clicking on the search result to pull in the file into the editor started maven off and running.
Maybe switch to intellij. Someone here was using that, but they left. Kind of hard to push yet-another-tool upon the team or be an outlier using a non-standard.
Some notes I took:
His slides are here
We have a case where people can submit multiple changes to their database records. Each change is handled separately with code like:
- from table where attribute=value
However, if the user happened to submit two changes to the same record the initial query will return the initially obtained object – NOT the updated value.
We fixed this by performing a refresh() once we get the object back from the hql query.
Met someone that works at a cloud software company. The company promise is that your infrastructure xxxx problems are handled by us in the cloud. Nice premise.
Reality is they are hobbling together a bunch of non-scalable apps written in a variety of languages together as a solution. The various pieces are re-developed (daily) into Java so that some kind of architecture can be established.
Yes, the original silo solutions are still being developed in their native language.
Yes, the amalgamating team has a daily run around to see what happened and ‘fix’ it.
Yes, there is an architect overseeing the lot and trying to herd these various individual teams into the same direction. (What’s that old joke about trying to herd a bunch of somethings?)
And, did I mention they are scattered across Europe so not in the same time zone or language.
Sign me up!
Saw this used on a job site to evaluate a position. Neat quick-and-dirty evaluation of a new position: A good company gets a 12.
The Joel Test
We have a fairly large database with years of stored procedures and views and the like.
Recently we tried to access a view to get a count of items right after we update the items list. Seemed simple enough: just flush so that all the ducks are in a row. Wrong.
Turns out the view is based on an indirect access to the item list we changed, so as far as Hibernate is concerned – nothing changed.
Another engineer had the breakthrough: evict the view thereby forcing Hibernate to re-establish the values. Ta Da!
Alternatively we could have mapped that indirect set that the view uses.