Dave

There's now an updated thread in the maths forums that describes some of the changes to the LaTeX rendering.

After a recent update to our forum software, typesetting equations on SFN has changed a little bit. Although we are still using LaTeX, for a variety of reasons, we've elected to shift over from our custom-written LaTeX generator to the excellent MathJax library, which will take your equations from post text and render them in your browser. Much as before, the idea is that in your post, you surround equations with special characters, and MathJax will convert the contained text into an equation for you. There's two types of equation that you can typeset: Inline math is displayed in the flow of a sentence, such as \( y= x^2 \). This example was produced by using the text \( y=x^2 \). Note that we do not support $ signs as most LaTeX users would be familiar with, since this occurs too frequently in text. Display math breaks up a paragraph and can be used for typesetting larger equations such as \[ y = \int f(x) dx. \] The text then picks up afterwards. This example was produced by using the text \[ y = \int f(x) dx \] , which we note is exactly what one would type in a usual LaTeX document. For reference, the old guide is still available and has a number of useful examples for those getting started. Finally, please note that for legacy posts, the old [math] [/math] tags will still continue to work and these will display equations as inline. However it's likely that older posts may look different to the way that they did before.

Just a quick note: if you want a single line break (instead of a paragraph) hit "shift+return" instead of just return. Additionally, for LaTeX, you can use the math BBCode as before, or you can use more standard notation to add display equations, e.g: \[\mathcal{F}(k) = \int_{-\infty}^{\infty} f(x)e^{ikx} dx \] produces \[\mathcal{F}(k) = \int_{-\infty}^{\infty} f(x)e^{ikx} dx \] I'll try to update the LaTeX guide as soon as I get a chance to reflect these updates.

One additional note from me: we no longer use our older LaTeX rendering system in favour of a browser-rendered MathJax, for various technical reasons to do with the way that posts are parsed and stored in this new forum software. The syntax for using our typesetting is exactly the same as before, but you will probably notice some rendering differences, particularly in older posts. Do let us know if there are technical issues on this front!

Just to throw in my two cents as well: this was an update that we really needed to perform, primarily due to security concerns of the older software being outdated (for around 2 years now, in fact). We'll do our best to smooth over any quirks of the updated software.

We'll schedule the IPS4 upgrade before the end of the month - our test install ran pretty well, so we can go ahead once we get a free block of time. In terms of the problem above, I am not convinced it's us that caused the redirect. The offending code is not currently present, as far as we can tell, so I would be more inclined to think it was the public wifi.

Hi all, Apologies, once more, for the downtime today (around the last 6-7 hours). Unfortunately, our server hosts experienced a major power outage this afternoon and the server did not automatically restart when the power came back up. Everything should now be back up, but please let me know if you encounter issues. Dave

Hi Dan, thanks for getting in touch - we had spotted this a few weeks ago, but the template cache did not get rebuilt so it has been lingering on a few pages. I have now rebuilt the caches and removed the offending code. It seems that there is some unknown attack vector, we believe inside IP.Board 3, that is allowing this to reoccur, since there are no other server infarctions and no out-of-place or different files from the original IPB installation. We're scheduling an update to IPS4 which should hopefully permanently eliminate this issue, but the update affects quite a bit of the site, so we have to do a little planning first.

Okay I think I tracked this down -- let me know if this is now fixed! Hope you're all having a good Christmas & Boxing Day

Thanks guys - will take a look this morning.

Hi all, Apologies for the extended downtime today. It was necessary to do a server upgrade and I choose one of our least busy times -- UK mornings, on Christmas Eve. Unfortunately before I could finish the update I became a bit ill so had to postpone the last parts until later this evening. Everything should be back in order but if not, let me know. Cheers, Dave

Apologies if you were caught in the ~20 minute outage that came rather suddenly today. There was an important security update that needed to be applied, and whilst there was a lull in activity I decided to capitalise on it! Please let us know in this thread, or in the Suggestions forum, if you encounter any problems.

Thanks for pointing out these issues guys and apologies for the lack of response. Unfortunately, IPB is a bit finicky when it comes to giving our LaTeX plugin non-htmlized text, so presently there's a bunch of hacks that need to be made in order to get it to work. Probably there are some characters missing from the table we use to do a search-and-replace. I think a better long-term solution for math typesetting is probably to utilise MathJax, which does away with the images and with it, a bunch of headaches such as baseline alignment for inline math. However, there is no IPB plugin for this at the moment, so will require me to do a little development in some downtime, which is pretty hard to come by at the moment!

Apologies for the slightly extended downtime, there was a slight misconfiguration in our networking interfaces which hadn't been spotted. Everything should now be back online, please let me know in this thread if you are experiencing any issues!

In order to respond to some security advisories, I will take the server offline at around 9am tomorrow (GMT+1) for approximately 1 hour to perform the necessary updates. During this time SFN will be unavailable. I will post here again once the maintenance is complete.

As far as we can tell, the problem has been fixed. It was a little hard to trace since it only appeared infrequently (roughly every 2 in 70 or so page requests according to Google). I will post a further update later as to the probable cause, but want to discuss the matter with the forum developers first. In the meantime we will keep a very close eye on the situation and await a review from Google.

Thanks studiot for the update. After a bit more searching we did identify the problem and have rectified it. Let us know if you see it again. Hopefully this should be sorted in the next few hours from the standpoint of Google and Safari/Firefox warnings.

Dear all, We received notification from Google a few hours ago that some malware had been injected into some of our forum pages, leading to a warning being displayed whenever SFN is listed in Google search results. Both myself and Capn have investigated this issue and have found no evidence of this so far -- it may be the result of an IP.Board vulnerability that we have just patched. Additionally, we have checked servers and done our best to ensure the site is safe. The issue has been bounced back to Google, and hopefully this should be resolved within the next few hours or days. I will post again once we have received an update.

Just a quick post to confirm that the server is now back online and the maintenance window was successful.

To install a new piece of hardware, I will need to take the server offline tomorrow morning (25th April) at 10am BST (9am GMT, 5am EDT) for approximately 15 minutes. I will update this thread once the upgrade is complete.

It completely depends on what your function is. Bear in mind that if you can't write a closed form for the inverse of the function then this approach won't really help unless you know something special about what happens when you apply a polar transformation to its inverse. It is entirely usual for the inverse of a function to exist but not be expressible in a closed form.

It's been a while since I looked at this, but I'm not sure the phrase "formal power series of a function" is well defined. You can calculate the power series of a (real/complex) function, which will generally be valid within some radius of convergence. That power series can also be viewed as a formal power series, in which we don't care about things like convergence, by choosing an indeterminate variable and the same coefficients as the power series. They won't be 'equal' because we (obviously!) don't know what type of variable the indeterminate is.

Ah yes, there does appear to be an issue there. I will address that momentarily. In the meantime, you should be able to access your blog directly at http://blogs.scienceforums.net/Tridimity/ Quick edit: this should now be fixed.

One nice feature for those amongst us with retina (or high-DPI enabled) monitors: LaTeX will now render at a much higher DPI for those displays, meaning that readability has been [imath]\text{greatly improved}[/imath]. (Note this won't show up on quick-edit yet, you need to refresh the page.)

We're back on the new server! Let us know if you find problems.