1) Contact Citrix support about some missing details in their SDK (Dynamic Virtual Channels were not loading properly on XenDesktop, and the SDK docs just said "Use the Microsoft APIs for registration")
2) Citrix ponders about the issue for a few months. I write a full DVC plugin for them to test things with, because they don't seem to have anyone in-house who knows their own tech stack
3) Citrix finally declares that dynamic virtual channels are currently broken. They might get fixed in a future release, but they're not on the active roadmap.
4) Citrix is still releasing frequent updates to the SDK, advertising DVC support. NB this is not just a small typo in the docs, they have a long chapter dedicated to the topic ( Citrix Dynamic Virtual Channel Protocol at https://github.com/citrix/receiver-for-windows-virtual-chann... ). The feature is very thoroughly documented, it's just...not actually implemented.
I can't imagine how many developer hours get wasted yearly with people trying to get virtual channels to work, when in reality they're just flat-out not supported.
That largely mirrors my experience with IBM's DB2 Connect software on Windows ~20 years ago. Follow the install documentation to the letter, it doesn't work. Contact support.
Months of follow ups later and:
IBM: "this isn't a supported configuration."
Me: "But, it's documented as a supported configuration."
IBM: "It has been removed from the latest documentation."
Me: "And when was that published?"
IBM: "Tomorrow"
Even after getting a DB2 Connect install working, it was a nightmare then dealing with their compatibility flags, which changed with every fixpack. Want feature A to work? Set the flag for it...but, you'll lose features B & C if you do that...
It's bizarre how companies reach a certain size, and they seem to lose all developers. There's the clueless helpdesk and the marketing/sales org, but nobody seems to be writing code / fixing issues anymore.
In my experience, contacting a 10-person company about a bug means that a dev is going to look at the issue in a day or two. Contacting a 10 000 person company about a bug means that a dev will never see the bug - unless they happen to be browsing Reddit and see someone complaining about it
Ubuntu was to be introduced into the company. The team was 1 developer and 4 managers. When we asked for collaboration with the developer the manager kept saying "we don't do this" "it's not a supported configuration".
The developer defied his direct superior and offered collaboration and as a result was removed from the project.
Now the project staff was 4 managers and 0 developers.
The manager claimed he was doing this to protect his developer from overwork, but really it was just about his own ego.
A lot of the projects I saw there were basically 1 engineer and a couple of managers of some sort. Security manager, Project Manager, Portfolio Manager, some medium management and some other people. And then they would constantly say that they don't have resources to fix things. Well yeah, because you have 1 engineer for 10 projects, but 4 managers per project.
Building new things is seen as business risk, I reckon.
This is why startups or VC backed things can explode, and innovate, but once you IPO, innovation is not what’s important - it’s about perspective I guess.
When the death spiral of anti-innovation starts, people such as talented developers start to drop off, and in the end you’ll have a slow, bureaucratic mess of an org, sustained by the inertia of an already existing customer base.
“How about we yank up them licensing fees?!”, or my favorite “Let’s rename a product suite!” (Citrix, as a case in point, is notorious for this…)
The developers are there but they are so far removed from the customer experience that it seems like they aren't. Large companies don't like developers talking directly with customers because they tend to speak frankly and honestly.
It's actually quite normal. Notice how most things posted on HackerNews are neither News, nor about Hacking.
It's normal, that things morph into some lowest common denominator 'thing' that makes people that have any taste shudder on the inside. It doesn't make most everyone else shudder, which is why it is almost inevitable, reduction to the lowest common denominator, of everything.
Why a message queueing technology is branded under 'Websphere' I have nooo idea, but it turned out to be a nightmare of various configurations jammed together and not working well.
Tibco Enterprise Messaging (EMS) on the other hand is a dream to work with IME.
I wonder if some large account acquisition somewhere somehow wound up contingent on implementing this functionality... but technical due diligence during the integration process identified that it was absolutely irrelevant... and it was subsequently declared that it was easier to just pretend the feature exists to keep the middle-manglement happy because the account was important.
I interviewed at Citrix 20+ years ago for my first SWE job out of college. The first phone interview went well, but on the second one I was surprised to find the interviewer was asking the exact same questions, word for word. After a couple of them I mentioned it, thinking maybe this was a mistake and the interviewer said that wasn't the case, and then proceeded to ask me the rest of the same questions. After I told them the answer to the next question before they asked it there was a real long pause on their side and then they asked the remaining questions in a different order. Needless to say I wasn't as excited about working there after that, but they never called me back either.
I was confused about Twitter's business model from the very start. But, I didn't think Uber and Lyft made a lot of financial sense, either :P
Usually what you'd do is pump the platform full of ads and then offer an ad-free option, or make it into a subscription service outright, but I don't think twitter has enough of a moat for the latter and won't be successful on the former.
Uber has only ever posted one positive quarter (Q3 2020) and has gone back to net losses in Q4. It doesn't really make financial sense.
Even the whole "they will be profitable once they have a monopoly" idea is overblown IMO. In Toronto were I'm located, we have a local taxi company (the largest one, Beck) who has an app that AFAICT has all the same features as Uber's core ride hailing app. Clearly Uber's core tech in ride hailing is not that much of an advantage...
Uber would've definitely be positive if it wasn't for COVID. Where I am, People use it massively to get to/from parties, that basically died for the entire 2021... The local Uber branch said there was less than 1/10 rides that year...
So in conclusion, I'd wait before saying what has/doesn't have financial sense.
> In Toronto were I'm located, we have a local taxi company (the largest one, Beck) who has an app that AFAICT has all the same features as Uber's core ride hailing app.
Well, and where's the big difference? Why your local app can work and Uber can't?
Your local app won't work in Italy, whereas Uber does, and works consistently and with the same credit card I use at home. IMHO it's cool.
My only experience with Citrix is their GoToMeeting app/service. It's probably the most aggravating meeting software I've ever used. From almost everyone else's comments here, GoToMeeting seems to be their best piece of software.
GoToMeeting is just as bad under LogMeIn as it was under Citrix. It’s barely changed in years so I assume no resources are being invested and they’re just milking the revenue until it eventually dies.
Used to work at Citrix for three years as a Product Manager. I empathize with most everyone's experience that they shared- at the tail end of my tenure, the company prioritized the sales teams over the engineering staff and unfortunately it showed within the product.
I think when set up properly, Virtual Apps and Desktops is killer for having good access to desktops and at the moment is a good extension of Azure Virtual Desktop. However, it did feel like we were running the product with a skeleton crew. As a result, documentation fell by the wayside along with other matters. I guess complete PE takeover of this sort of company is inevitable.
Worked for a company that Citrix acquired a few years ago. They seemed to have went all in on something called Kepner-Tregoe. I left shortly after my week long training course, which I didn't get much out of other than a stay in a nice hotel. As I was leaving they were downsizing and one of their techniques was to match 2 people together and say "We will be comparing your performance against each other for the next 6 weeks, after that, the worst performer will be fired". Pretty awful.
So little content in the article about why this was needed. The argument about improving analytics for Citrix products using Tibco products seems extremely weak. I don't see much if any opportunities here for any combined improvement at the tech stack level. Where there is any opportunity there is redundancy like gateway functionality.
All the best to those working there and trying to figure out what next.
When I saw the headline I started laughing. I cannot for the life of me understand how these two companies make sense to glue together. Is Vista hoping it's like a bundled deal in the grocery store? Nobody wants to buy this particular brand of chocolate or this particular flavor of ice cream, but if I put it in a plastic bag and sell them together as a "weekly special" somebody will buy it???
My guess: Tibco has been owned by Vista for 6 or more years. They trust the management and they'll use them to reorg Citrix - reducing finance, HR and marketing expenses.
Good for them; Azure Virtual Desktop and Amazon Workspaces are both pretty mature offerings that absolutely blow Citrix out of the water on cost, performance and functionality. It’s one of those rare use cases where autoscaling in the public cloud with some lightweight management tools is an unreasonably effective solution by nearly every available metric.
I can't speak for Amazon Workspaces, but Azure Virtual Desktop by itself is a terrible experience. Citrix provides a pretty large value-add for AVD in particular: FAS + Azure AD integration means a user just needs to sign into their local (Windows) system and that's it, they're automatically signed into Storefront or Citrix Cloud through their Azure AD PRT and then FAS signs them into their VDI system when they connect. Add on HDX, support for things like FIDO2 redirection, etc. and I would say the exact opposite of you.
I mean, using it's fine too if you're not like so many organizations that vastly over-provision their infrastructure. I did a PoC with Citrix Cloud (Azure, Windows 10 multi-session) prior to us rolling it out for a family practice location and was rather impressed with the performance when using Standard_D4a_v4 instances - round trip latency metrics available in Citrix Studio showed around 60ms from keypress to update and my eyeballs would back that up.
I've also used VDI setups that were miserable with heavily over-provisioned CPU resources and 2-vCPU's per VM. The experience really is dictated by the hardware you allocate.
For all those ranting a only partial rant: Citrix, I did use it once kind of without anger at it: to run an unpatched IE8 instance to run an even older unpatched ASP (not .net) application (to run some ancient ERP software from BAAN) which I did get angry at.
I worked on Citrix-remoted VDI machines at two previous jobs. The first one was Windows 7 on dedicated hardware with some kind of a custom client on top of Citrix, and the second one was Windows 10 virtual machine (no GPU) and a WYSE terminal as a client.
Windows 10 seems to be not remotable at all due to the fact that everything is rendered in 3D space, so the only way to send data to the client is by sending bitmaps. Plus, there is added latency due to the fact that you have to cache a frame or two to optimize compression. This made it impossible to work, those dreaded delays in key presses, oh God. As I heard from a guy at the top, adding a GPU to the server helps a lot. But no corpo would do that en masse. It will for sure cause an exodus from this VDI setups in the nearest future. Maybe apart from the cheapest people in India.
I have fiber with 6ms ping to our windows machines, RDP connection. I've tried a few times to get these VDI setups close to matching what I have now and I can't. Either RDP is amazing, naked hardware is amazing, or VDI is a dog. I'm willing to pay.
I still haven't found anything better than RDP. Back when Windows was my primary work machine, I could bring it home, plug it into power, shut the lid, toss it up on a shelf and still RDP into it, and use all of my monitors. And it ran as good as local. Still haven't found anything that compares to it in any way. VNC always just gives you the resolution of the host's desktop. Plus it's always just some screen capture into a video codec.
Ok, well back in the day, I would say X11 was actually far better than that. Over dialup, it would blow away RDP performance-wise.
TIBCO has a lot of Analytics and Data products. One of their acquisitions, called Spotfire, is actually a really good product.
Over the last few years they've also acquired a lot of Data Science companies, but it always felt like they were bargain shopping and acquiring companies nobody else wanted to buy.
Honestly, I pity Spotfire. IME it's the best out of all the products in that space (e.g. Tableau, PowerBI, etc) and is much better for actually Analyzing data in comparison to other products it competes with. But they got acquired by TIBCO, who basically did a great job of losing the mindshare race against Tableau. I feel like some other company would have done a better job with Spotfire.
Citrix does desktop virtualization. Back when I was involved with using their stuff (20 years ago), it was a lot faster for global users to connect to our Citrix server to run client/server apps than to run them over the WAN.
Tibco does/did middleware for n-tier applications. Back when I used it (again 20 years ago) it worked, and was relatively painless.
Tibco is short for “the information bus company”. They practically invented Enterprise integration in large businesses, esp stock exchanges and banks, with a need for high reliability. These days Tibco is largely considered “legacy”. I’m surprised that VCs would see value in them apart from their Rolodex.
Tibco is owned by Private Equity, not VCs. Very different type of companies. VCs own a company on its way up, PE onws its way down. Some times a company is sold immediately from VCs to PE. In some rear occasions a company owned by PE can go public: e.g. Dynatrace, TeamViewer.
Yup. When I was a junior engineer, I was working for a bank and the various teams were all blaming each other about high latency in the stack, which involved Tibco Rendezvous. Basically, a way to distribute market data over multicast. They blamed Rendezvous, they blamed the network hardware, they blamed everything except their spreadsheets. (Yes, the tech stack at the time was spreadsheet <-> spreadsheet over multicast.) Anyway, I wrote a small Perl application that measured the end-to-end latency (when market data changed, vs when the published output changed) in addition to network induced latency (basically echo packets; write something out, see when it came back over the link). What the data showed was that the data processing spreadsheet introduced tons of random latency, and the network was totally fine. I sent that up the chain and the bug in the spreadsheet was found almost immediately, solving a multi-month blame game in hours.
I worked for a Vista-owned company for about 8 years. We were a division of a larger company spun out and sold to Vista. We eventually merged with a couple of other Vista purchases.
The initial change I remember was that everyone had to take a combo personality/IQ test, the DPAT (dynamic personality assessment test?). It looks like they have replaced that with something called a CCAT now, but I'm sure it's essentially the same thing. The results of your test were not shared with you, and IIRC we were told that the results were available only to the CEO and their executive assistant. Candidates for hire had to take this test too, and if they didn't pass we couldn't hire them; it was challenging to hire people in our area. The general feeling was that no one applying to work with us was passing the DPAT, although I wasn't personally involved in any discussions where someone who would know that (CEO and EA only, remember...) told me so.
Some time after this test - maybe a few months? - we had a large layoff. I don't know how many people it was, or how many people we had, actually. I believe the entire company had less than 2500 people, and R&D was at most a few hundred. But they brought in security guards to patrol the office for a week or two during and after the layoff. Vista wanted us to hire "HPEL"s - high potential entry level. We seemed to have layoffs every 12-18 months after that. We were always told that change was a constant, things were looking up, we don't anticipate more layoffs, did you know we are hiring!, etc.
We were paid, and not tortured or anything. But pay was always below-market, and while my memory is fuzzy on the exact sequence of events, it seemed like we were changing 401k and health providers every couple of years. All hands meetings where people complained about compensation and such would get answers like "we continually evaluate industry conditions and remain competitive in the market etc. etc. etc." Leadership was not constant, and from my position it felt like we never remained committed to anything long enough to succeed. The meme where a baby runs down a hall, encounters something alarming at the corner, and then turns and starts to run back? It felt like that most of the time.
The other thing that I thought was fishy but not surprising is that we developed business relationships with other Vista companies. We used Tibco, BigMachines, Marketo, Ping Identity, and maybe WebSense. We had training stuff from SumTotal. I don't remember there being an official policy around this - although Vista had their 'Vista Operating Procedures' playbook for acquisitions to execute - but it did seem like the invisible hand was pressing its invisible thumb on the scale somewhere.
Morale was mostly terrible. I know it wasn't just me that felt that way, but I don't know how much of it was due to Vista vs. something else about our company. I do feel like PE acquiring you means you're not successful, that they think they're going to lean you out and flip you for a profit. So I would be worried if I were at Citrix, and if my own experience and observation is relevant I would suggest dusting off resumes and at least being prepared to move.
Spent ten years at a Vista company that I won't name, half before acquisition and half after, and all of this pretty much tracks. (Maybe not the security guard part, but I can confirm that when you were laid off, you were escorted straight out the door.)
I was a manager and hired people, and I never had a candidate who didn't pass the CCAT. But I'd have to qualify that with "a native speaker of English" and "in the US." I never hired in India or elsewhere overseas, but I heard from other managers that this could be a challenge, which makes me think the CCAT was a convenient way of weeding out "non-culture fits."
As far as layoffs and CCATs, they layoff decisions were always a few steps above me in the org chart, but what I observed was that the last couple of rounds were very directly related to employee age. NOT tenure, although being there for 20 years meant you were on the chopping block to be replaced by a few HPELs. But I saw people who were 50+ and only around a few years get cut. Maybe it was due to salary, maybe not. Officially, it was due to "a formula" that included "several factors" because of course basing layoffs on age is illegal.
Annual reviews were a bit bogus, because they were 1-5, with 1 being you were actively involved in a felony against the company, 2 being you were on a PIP, and you couldn't give a 5 to anyone. I would give a scattering of 3s and 4s, and then my boss's boss would say I had too many 4s, so effectively everyone was a 3, which doesn't do much for the "several factors." (It did play well into massively underfunding the bonus pool, though.)
The only other part missing from your story is how between every layoff cycle, there would be a giant rumor that Broadcom or Oracle or someone was going to buy the company for the rolodex and real estate, and fire everyone they couldn't offshore.
With exception of the security guards, your experience is spot on with one of my experiences working for (and through) a Vista acquisition.
To me the morale killer was the constant change. HR constantly opening positions and then freezing hiring before we could hire somebody, resulting in so much wasted time interviewing people. We'd work on big project for a while, and then leadership would change directions and all that was a waste. By the time the 5th person on my team left, it was hard to not feel like everyone was graduating but I was the only person who failed in my senior year and was left behind.
Same with Jaspersoft, I used to enjoy working with their technical guys and went to a few conferences, but then it disappeared into the larger Tibco bucket. Tibco did spend a bucketload of cash on a hotel for their conferences though, very flash but I didn't get much out of it.
I hate citrix. It’s such a shitty aggravating user experience. Sometimes in the middle of the night with no load it almost works ok but it’s never good.
In my experience this is almost always because your IT department budgeted terribly for their deployment and either didn’t get GPUs for graphics or is overextending their setup.
VDI should probably cost as much if not more than comparable workstations, but it’s so much more convenient in the best case.
That’s always the pitch that it’s not setup correctly but if most users are experiencing it in the incorrect setup then maybe it’s just the product. Like I said in no load it can be ok but never good. Sometimes the security might be a good trade off but only if users interact with the vdi rarely during their work day
A lot of times the issues are disk IO, 500 desktops all booting up and end-users opening applications at around 9:00am is a surprisingly huge amount of IO.
spot on. i did a brief stint as a xendesktop admin. it’s quite good if you throw big hardware at it, but so many places were cheap and overprovisioned the piss out of their underlying ESXi hosts (or did no resource scoping because surely the folks in accounting that mostly use Excel have the same IO needs as developers compiling multi gig projects in .NET) and they absolutely were not paying for PCoIP and/or GPUs (which you really only needed if low latency VDIs were a necessity)
The installation wizard for workspace and realtime engine on macOS installs 6 or more background services as root which you don't actually need if your provider just gives you an ICA file to login.
I took a python class in a Citrix environment and It was incredibly aggravating. Watching code appear seconds after you type it fills you with so much rage. It’s like zoom fatigue before zoom
1) Contact Citrix support about some missing details in their SDK (Dynamic Virtual Channels were not loading properly on XenDesktop, and the SDK docs just said "Use the Microsoft APIs for registration")
2) Citrix ponders about the issue for a few months. I write a full DVC plugin for them to test things with, because they don't seem to have anyone in-house who knows their own tech stack
3) Citrix finally declares that dynamic virtual channels are currently broken. They might get fixed in a future release, but they're not on the active roadmap.
4) Citrix is still releasing frequent updates to the SDK, advertising DVC support. NB this is not just a small typo in the docs, they have a long chapter dedicated to the topic ( Citrix Dynamic Virtual Channel Protocol at https://github.com/citrix/receiver-for-windows-virtual-chann... ). The feature is very thoroughly documented, it's just...not actually implemented.
I can't imagine how many developer hours get wasted yearly with people trying to get virtual channels to work, when in reality they're just flat-out not supported.