Should Tumblr care? David Karp tells users who complain to "go away"

William's email was meant to make people feel bad. Look at the first sentence: "I am writing this email to you to express my recent disappointment with Tumblr". Instead of feeling bad, David got irritated. Thats the second best choice (don't let people make you feel bad for building gnarly web infrastructure, ya know?).

The best choice would be to ignore the email and let someone without emotional attachment handle it. Sometimes, though, you can't fight the urge to slap down the Wah-wahs.

The original article, by the way, is just trying to paint David Karp as an arrogant, unlikeable person. There are mounds of unreferenced personal attacks in there.

Its about implementing strategy for that.

"And this right here folks is why a startup is so extraordinarily stressful, why PR agencies exist, and why a squeaky wheel gets the grease." - Ramanujan http://news.ycombinator.com/item?id=2151989

I have engaged in many email interactions with the Tumblr staff and outside of a few questionable missives, encountered nothing but gracious and welcoming responses. Unlike other cloud services like Google, which is merely interfacing with a Python script HTTP request.

That said, Tumblr is straining to keep pace with growth, and the service is sporting cracks and leaks that it do not appear to be getting resolved…

* …the well known cited issues about availability, service outages are still experienced on a daily basis, even if it is just for a few moments. Confounding that the Dashboard might be available but accessing your site page via browser URL generates an error. Or when publishing from the queue, an error is generated even though the article is published -- it just is not removed from the queue, so you have manually delete it from the queue after publishing (reckon there is no DB transaction on the set of CRUD operations).

* …regarding the queue, it is a great feature if it is functional, but since last fall, it is not really functional. It sort of works now, mostly as a holding bucket for me -- I realize in the eyes of Tumblr Support, it works, outside of some outages (which have been measured in days length, not hours), but it is just not reliable enough to depend on. Also, in the wake of the refresh, the UI was badly butchered -- relying on AJAX drag and drop that makes it impossible to move items to the top or bottom of queue, plus some other befuddling choices that make me question whether the developers making the modifications ever actually use this feature.

* …the rich text editor continues to deteriorate (you have the option of editing in rich text, raw html or markdown), there is a glitch now that if you click on the "Edit HTML source" button, it just "hangs" on a blank window. But even when working, it inserts all sorts of extraneous HTML entities and spans. Have switched back to the Markdown editor, but it's an incomplete implementation.

* …the schizophrenic UI nature of accessing Tumblrs via the dashboard or the site pages -- there are sets of operations that you can do on the dashboard (reply) that cannot be performed from the site page.

* …the API access (for creating/updating/dashboard requests) is insecure and really has not been enhanced or upgraded in years.

* …there is a Mac app (no Windows or Linux) to export your Tumblr, but it has failed to export every time I've attempted.

Going to stop there, as I don't wish to bang on Tumblr -- I know they're working hard to address the performance issues. I just hope that it is not a rabbit v. tortoise race.

However, the quoted email definitely was.

(So I guess spolsky referred mostly to that one, not just the article around it.)