I know it's not a popular bullet point, but until Google Docs is 100% compatible, complaints like this don't have much grounds to stand on.
The difficulty: 100% compatibility is impossible even in theory. All that Microsoft need do is push a patch that adds or changes some minor feature, and now its competitor is no longer "100% compatible." So a Microsoft troll can always argue the "100% compatible" point, and win it by moving the goalpost. (I don't argue that you are a Microsoft troll, by the way: merely that you are advancing the argument that a Microsoft troll would advance.)
Of course, in the real world, the goalpost does move, because Microsoft is a monopoly and has captured the government. Microsoft products cannot interoperate perfectly even with themselves (ever try to open a 10-year-old document?). Nevertheless, failure to interoperate perfectly with Microsoft products dooms all competition to irrelevancy as "non-mainstream." Any non-Microsoft software, to be successful, has to address a problem for which Microsoft does not have even a bad solution. For instance, software to simulate the operation of a chemical plant might stand or fall on its technical merit. But even that might fail, because it likely has inputs and outputs that are tables of numbers, and therefore will face "100% compatibility" issues with Excel.
Given that no government at this point can intervene without crippling itself - governments run on PowerPoint, and Microsoft has them by the short hairs - I don't see the situation as changing very soon.
I didn't test PHP, because I never use it.
I did, however, do the responsible thing and ask myself, "does code that I maintain have a similar bug, and does its test suite validate it?" I checked the Tcl test suite, and discovered that it lacked a test for the case. I added test cases for input and output conversion of the largest positive subnormal, the smallest negative subnormal, the smallest positive normal, and the largest negative normal. (All the test cases passed, so committing them ended what I had to do.)
Tcl, too, has its own conterparts to strtod and sprintf("%g"), because it depends on float->string->float being 100% lossless and on having the string representation of a float be the shortest string that reconverts to the given number. It's quite tricky to get right; this particular corner case is the tip of the iceberg.
It's always human error:
The careful text-books measure
(Let all who build beware!)
The load, the shock, the pressure
Material can bear.
So, when the buckled girder
Lets down the grinding span,
'The blame of loss, or murder,
Is laid upon the man.
Not on the Stuff - the Man!
- Rudyard Kipling, "Hymn of Breaking Strain"
A system using proper fractions can actually get exactly the right answer every time OR it will overflow and we will know for a fact the answer isn't exact.
What theory of numeration are you using, that has all numbers rational? I'm sorry, but even the humble square root is something I don't want to give up, to say nothing of transcendental functions. The theory of exact arithmetic on the reals is not all that well developed. Bill Gosper makes a start, and a handful of researchers take it somewhat further, but actually using exact arithmetic for everything you'd want to do remains a mirage.
Well, it's at least a start. Now instead of having one case against defendants across the whole US, the lawyers have to refile in all 92 district courts, multiplying their expenses considerably. (Particularly since a lawyer who files pro hac vice in many of the districts must retain local counsel.) Moreover, the defendants will be sued in their home districts. They don't need to retain counsel and appear in DC to contest their cases. (If a distant court finds a way to assert personal jurisdiction, you're well advised either to prosecute an interlocutory appeal or settle, because the expense of simply making all the required appearances in the distant state will be ruinous.)
This was also a mandatory first move on the defendants' part. Once you've filed any paperwork with the court without contesting its jurisdiction, you've lost your chance. If you've been sued in the wrong place, you get only one short opportunity to tell that to the judge. After that, you're deemed to have acquiesced on the point. That's unbelievably harsh against pro se defendants, who may not have even been able to find a lawyer (especially since the notice served against them states that they need fo find one in the distant state) by the time they're required to file an answer, but it's the law.
So it's not a major victory, but it would have gone considerably worse for the defendants had the judge asserted personal jurisdiction against them.
Any circuit design must contain at least one part which is obsolete, two parts which are unobtainable, and three parts which are still under development.